Methods for characterizing and manipulating protein-protein interactions - Patents.com
The AlphaSeq platform facilitates high-throughput identification of compensatory mutations in protein-protein interactions, enhancing our understanding of binding kinetics and aiding in the development of novel antibodies by comprehensively screening libraries of mutant polypeptides.
Patent Information
- Application Number
- JP2022573200
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2020-06-01
- Filing Date
- 2021-06-01
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2041-06-01
AI Technical Summary
Existing methods for characterizing protein-protein interactions, such as epitope and paratope mapping, are limited in their ability to efficiently and comprehensively identify compensatory mutations between protein binding partners, which are crucial for understanding binding kinetics and differentiating novel antibodies in crowded IP spaces.
A method involving exhaustive site-saturation mutagenesis and high-throughput screening using the AlphaSeq platform to create libraries of mutant polypeptides, allowing for library-by-library screening to identify compensatory mutations between protein binding partners, particularly antibodies and antigens.
Enables rapid and cost-effective identification of compensatory mutations that enhance or restore binding affinity, providing unique insights into protein-protein interfaces and aiding in the development of novel antibodies and therapeutic applications.
Smart Images

Figure 0007588659000005 
Figure 0007588659000006 
Figure 0007588659000007
Abstract
Description
[Technical field]
[0001] Related Applications
[0001] This application claims priority to U.S. Provisional Patent Application No. 63 / 033,176, filed June 1, 2020, the entire contents of which are incorporated by reference herein. [Background technology]
[0002] background Epitope mapping is an experimental process that characterizes the identity, amino acid composition, and conformation of an antibody's binding site on its target antigen. Epitope mapping can be useful in the discovery and development of novel therapeutics, vaccines, diagnostics, among other biomedical applications. Epitope mapping can also be useful, for example, in ensuring intellectual property (IP) protection for novel therapeutic antibodies. Comprehensive characterization of the amino acid identity and conformation of epitopes of novel antibodies can help define the novelty of the antibody, the non-obviousness of the antibody, and provide the written narrative support required for the disclosure of novel antibodies. A crowded IP space, for example a therapeutic target against which multiple drugs already exist, requires the ability to differentiate novel antibodies against the same target from previously disclosed antibodies.
[0003]
[0003] Similarly, paratope mapping is the characterization of the properties of an antibody that confer its specificity to its antigen, such as amino acid composition, charge, and three-dimensional conformation. Thorough characterization of antibody-antigen interactions by both epitope mapping and paratope mapping is useful for understanding the mechanism and kinetics of specific binding between antibodies and antigens, and can be used to gain structural insight into the binding interface. Methods for epitope and paratope mapping include array-based oligopeptide scanning, site-directed mutagenesis mapping, high-throughput shotgun mutagenesis mapping, and cross-linking-coupled mass spectrometry, among others.
[0004]
[0004] More broadly, characterization of the binding dynamics at the interface between any two specifically interacting proteins plays a role in various biomedical applications. The methods disclosed herein can provide high-throughput characterization of the specific interactions at the interface between two protein binding partners. In certain embodiments, the methods disclosed herein use a combination of exhaustive site-saturation mutagenesis and high-throughput screening to comprehensively characterize the interactive surfaces of two protein binding partners simultaneously in a cost-effective rapid assay. The methods disclosed herein can be used to characterize any two protein binding partners, for example, for simultaneous epitope and paratope mapping of an antibody and its antigen. Summary of the Invention [Means for solving the problem]
[0005] Overview of Exemplary Embodiments
[0005] The foregoing general description of illustrative implementations, and the following detailed description thereof, are merely exemplary embodiments of the teachings of the present disclosure and are not intended to be limiting.
[0006] In some implementations, the present invention provides a novel method for identifying compensatory mutations between two protein binding partners, comprising: providing a first library of protein-binding partners, the first library of protein-binding partners comprising a first wild-type polypeptide and a first plurality of mutant polypeptides; providing a second library of protein-binding partners, the second library of protein-binding partners comprising a second wild-type polypeptide and a second plurality of mutant polypeptides; determining affinity measurements between each protein binding partner of the first library of protein binding partners and each protein binding partner of the second library of protein binding partners; identifying one or more pairs of protein binding partners having respective affinity measurements that differ significantly from the affinity measurement between the first wild-type polypeptide and the second wild-type polypeptide based on the affinity measurements between each protein binding partner of the first library of protein binding partners and each protein binding partner of the second library of protein binding partners; The present invention provides a method comprising:
[0007] In some implementations, the present invention provides a novel method for identifying compensatory mutations between two protein binding partners, comprising: providing a first library of protein binding partners comprising a first wild-type polypeptide and a first plurality of mutant polypeptides; providing a library of two protein binding partners, the library comprising a second wild-type polypeptide and a second plurality of mutant polypeptides; determining affinity measurements between each protein binding partner of the first library of protein binding partners and each protein binding partner of the second library of protein binding partners; and based on respective affinity measurements between each protein binding partner of the first library of protein binding partners and each protein binding partner of the second library of protein binding partners, (i) a polypeptide of a first plurality of mutant polypeptides, and (ii) identifying one or more pairs of protein-binding partners that comprise a polypeptide of the second plurality of mutant polypeptides; the measured affinity of each of the one or more pairs of protein-binding partners is significantly different than the respective expected affinity between each respective pair of protein-binding partners; The affinity expectation for a given pair of protein binding partners is a) an affinity measurement between a first wild-type polypeptide of a given pair and one polypeptide of a second plurality of mutant polypeptides of a given pair; and b) an affinity measurement between one polypeptide of a first plurality of mutant polypeptides of a given pair and a second wild-type polypeptide of a given pair; The present invention provides a method for calculating the
[0008] BRIEF DESCRIPTION OF THE DRAWINGS
[0008] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate one or more embodiments and, together with the description, explain these embodiments. The accompanying drawings are not necessarily drawn to scale. Values and dimensions shown in the accompanying graphs and drawings are for illustrative purposes only and may or may not represent actual or preferred values or dimensions. Where applicable, some or all features may not be shown to aid in the explanation of the underlying features. [Brief description of the drawings]
[0009] [Figure 1]
[0009] A series of charts showing the library-by-library screening capacity of the AlphaSeq™ method are presented. [Figure 2A]
[0010] Schematic diagram of two protein binding partners interacting in a complex, where the first protein binding partner can be an antibody and the second protein binding partner can be an antigen. Residues of both protein binding partners at the interface between the proteins are numbered. [Figure 2B]
[0011] Strength measurements per library are shown by AlphaSeq of interactions between protein binding partners. Site-saturation mutagenesis was performed at position 19 in one protein binding partner and at position 32 in the other protein binding partner. The inlay shows the interaction measurements between all amino acid single mutations at the two positions. [Figure 3A]
[0012] 1 is a graph depicting the interaction between two protein binding partners exhibiting orthogonal binding. [Figure 3B]
[0013] 1 is a graph depicting the interaction between two protein binding partners exhibiting receptor-specific binding. [Figure 3C]
[0014] 1 is a graph depicting the interaction between two protein binding partners showing ligand-specific binding. [Figure 4]
[0015] Illustrates the workflow of a library-wise protein-protein interaction screen using the AlphaSeq platform. [Diagram 5]
[0016] 1 is a plot of AlphaSeq protein interaction data showing antibody-antigen interactions measured on the AlphaSeq platform. [Figure 6]
[0017] Illustrated are the results of an AlphaSeq experiment in which eight antigen variants were screened against eight antibody variants, detecting and quantitating 64 interactions. [Figure 7]
[0018] 1 is a heat map showing the results of a screen of a PD-1 site-saturation mutagenesis library against wild-type pembro scFv (antibody). The residue distance between a given PD-1 residue and the nearest pembro residue is also shown. [Figure 8]
[0019] FIG. 11 highlights specific residues in the crystal structure of the PD-1 / pembrolizumab interface. [Figure 9]
[0020] 1 is a heatmap showing the results of a clean pembro scFv site-saturation mutagenesis library against wild-type PD-1 (antigen). [Figure 10]
[0021] 1 is a graph depicting the crystal structure of the PD-1 / pembrolizumab scFv interface. [Figure 11]
[0022] The structure of the PD-1 / pembrolizumab interface is shown. [Figure 12A]
[0023] 1 is a heat map showing pembrolizumab amino acid residues that were found to be particularly intolerant to mutations. [Figure 12B]
[0024] Model representing the crystal structure of the PD-1 / pembrolizumab interface and highlighting amino acid residues found to be particularly intolerant to mutations. [Figure 13]
[0025] 13 is a table of compensatory mutation pairs identified by AlphaSeq from a single assay. [Figure 14]
[0026] Affinity strength data for PD-1 and pembrolizumab mutations are presented along with a graph showing the crystal structure of the antibody-antigen interface. Some amino acid positions are at the interface but are highly tolerant to mutations. [Figure 15]
[0027] FIG. 1 is a diagram illustrating the ability of the AlphaSeq platform to detect compensatory mutations by measuring relative AlphaSeq signal in a library-by-library screen of pembro scFv (antibody) vs. PD-1 (antigen). [Figure 16]
[0028] Shown are plots of three pairs of mutant protein binding partners that display a signature of compensatory mutations. [Figure 17]
[0029] Two pairs of plots of mutant protein binding partners showing the signature of compensatory mutations are shown. [Figure 18]
[0030] Graph highlighting pairs of compensatory mutations detected by measuring relative AlphaSe signal in a library-by-library screen between libraries of pembro scFv (antibody) and PD-1 (antigen) mutations. [Figure 19]
[0031] 1 shows a method for epitope mapping by targeted mutagenesis. [Figure 20]
[0032] 1 shows an epitope mapping library-by-library screen using the methods disclosed herein. [Figure 21]
[0033] 1 is a heat map showing the results of a screen of PD-1 variants against a library of antibodies. [Figure 22]
[0034] 1. Enrichment / depletion heatmap showing the results of the screen per library. [Figure 23]
[0035] FIG. 1 is a schematic diagram showing protein binding partners, where compensatory mutations are identified between a first protein binding partner and a second protein binding partner. [Figure 24A]
[0036] 1 shows an epitope mapping library-by-library screen using the methods disclosed herein. [Figure 24B]
[0037] FIG. 13 is a heat map showing intensity data for 20 PD-1 variants harboring mutations at single amino acid residues and 20 pembrolizumab variants harboring mutations at single amino acid residues, or pairwise interaction data for a library of PD-1 mutations and a library of pembrolizumab mutations, with an expanded inlay showing all 400 protein-protein interactions measured by the methods disclosed herein. [Figure 24C]
[0038] Highlight specific pairwise interactions between single PD-1 mutations and single pembrolizumab variants. [Figure 24D]
[0039] Graph showing four pairwise interactions between wild-type and mutant PD-1 in combination with pembrolizumab. [Diagram 25]
[0040] 1 shows a first plot of expected and measured interaction strengths between two protein binding partners, and a second plot of measured versus expected interaction strengths between antibody-antigen protein binding partners assessed using the methods disclosed herein. [Figure 26]
[0041] 1 is a plot of the ratio of measured to expected interaction strength versus the distance between amino acid residues between protein binding partners. [Figure 27]
[0042] This is a three-dimensional model based on the X-ray crystal structure of the interface between PD-1 (antigen) and pembrolizumab (antibody). DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0010] DETAILED DESCRIPTION OF EXEMPLARY EMBODIMENTS
[0043] The following description in conjunction with the accompanying drawings is intended to be a description of various exemplary embodiments of the disclosed subject matter. Specific features and functions are described in conjunction with each exemplary embodiment. However, it will be apparent to one of ordinary skill in the art that the disclosed embodiments may be practiced without each of those specific features and functions.
[0011]
[0044] Throughout this specification, references to "one embodiment" or "one implementation" or "one implementation" mean that a particular feature, structure, or characteristic described in connection with one embodiment or implementation is included in at least one embodiment of the disclosed subject matter. Thus, the appearances of the phrases "in one embodiment" or "in an embodiment" in various places throughout this specification do not necessarily refer to the same embodiment. Furthermore, particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. Furthermore, embodiments of the disclosed subject matter are intended to cover modifications and variations thereof.
[0012]
[0045] It should be noted that, as used herein and in the appended claims, the singular forms "a," "an," and "the" include plural referents unless the context clearly indicates otherwise. That is, as used herein, words such as "a," "an," and "the" have the meaning "one or more" unless expressly specified otherwise. Furthermore, it should be understood that terms such as "left," "right," "upper," "lower," "front," "rear," "side," "height," "length," "width," "top," "bottom," "internal," "external," "inside," and "outside" that may be used herein merely describe a point of reference and do not necessarily limit the embodiments of the present disclosure to a particular orientation or configuration. Furthermore, terms such as "first," "second," and "third" merely identify one of the many parts, components, steps, operations, functions, and / or points of reference disclosed herein and similarly do not necessarily limit the embodiments of the present disclosure to a particular configuration or orientation.
[0013]
[0046] Additionally, the terms "approximately," "about," "approximate," "minor variations," and similar terms generally refer to ranges that include the specified value within a margin of 20%, 10%, or preferably 5%, in certain embodiments, and any value therebetween.
[0014]
[0047] All features described in connection with one embodiment are intended to be applicable to the additional embodiments described below unless expressly stated otherwise or unless the feature or function is incompatible with the additional embodiments. For example, if a desired feature or function is explicitly described in connection with one embodiment but not explicitly mentioned in connection with an alternative embodiment, it should be understood that the inventors intend that the feature or function can be developed, utilized, or implemented in connection with the alternative embodiment unless the feature or function is incompatible with the alternative embodiment.
[0015]
[0048] The practice of the techniques described herein may employ conventional techniques and explanations of organic chemistry, polymer technology, molecular biology (including recombinant technology), cell biology, cell culture, biochemistry, protein engineering and sequencing techniques, unless otherwise indicated, which are within the skill of the art. Such conventional techniques include bacterial, fungal and mammalian cell culture techniques and screening assays. Specific illustrations of suitable techniques can be obtained by reference to the examples herein. However, of course, other equivalent conventional procedures can also be used.Such prior techniques and descriptions can be found in Green, et al., Eds. (1999), Genome Analysis: A Laboratory Manual Series (Vols. I-IV); Weiner, Gabriel, Stephens, Eds. (2007), Genetic Variation: A Laboratory Manual; Dieffenbach, Dveksler, Eds. (2003), PCR Primer: A Laboratory Manual; Bowtell and Sambrook (2003), DNA Microarrays: A Molecular Cloning Manual; Mount (2004), Bioinformatics: Sequence and Genome Analysis; Sambrook and Russell (2006), Condensed Protocols from Molecular Cloning: A Laboratory Manual; and Sambrook and Russell (2002), Molecular Cloning: A Laboratory Manual (all from Cold Spring Harbor Laboratory Press); Stryer, L. (1995) Biochemistry (4th Ed.) W.H. Freeman, New York NY; Gait, “Oligonucleotide Synthesis: A Laboratory Manual” (Vols. I-IV); Practical Approach” 1984, IRL Press, London; Nelson and Cox (2000), Lehninger, Principles of Biochemistry (3. rd Ed)., WHFreeman Pub., New York, NY; Berg et al. (2002) Biochemistry (5 th Ed.), WH Freeman Pub., New York, NY, all of which are incorporated herein by reference in their entireties for all purposes.
[0016]
[0049] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. All publications mentioned herein are incorporated by reference for the purpose of describing and disclosing devices, methods, and cell populations that can be used in connection with the inventions described herein.
[0017]
[0050] The term "complementary nucleotides" as used herein refers to Watson-Crick base pairing between nucleotides, specifically nucleotides hydrogen bonded to each other by a thymine or uracil residue bound to an adenine residue by two hydrogen bonds, and cytosine and guanine residues bound by three hydrogen bonds. In general, nucleic acids include nucleotide sequences that are described as having a "percent complementarity" or "percent homology" to a particular second nucleotide sequence. For example, a nucleotide sequence may have 80%, 90%, or 100% complementarity to a particular second nucleotide sequence, indicating that 8 out of 10, 9 out of 10, or 10 out of 10 nucleotides of the sequence are complementary to the particular second nucleotide sequence. For example, the nucleotide sequence 3'-TCGA-5' is 100% complementary to the nucleotide sequence 5'-AGCT-3'; the nucleotide sequence 3'-TCGA-5' is 100% complementary to the region of the nucleotide sequence 5'-TTAGCTGG-3'.
[0018]
[0051] "Homology" or "identity" or "similarity" refers to sequence similarity between two peptides, or, more frequently in the context of this disclosure, between two nucleic acid molecules. The term "homologous region" or "homologous arm" refers to a region on a donor DNA that has a degree of homology with a target genomic DNA sequence. Homology can be determined by comparing a position of each sequence that can be aligned for comparison. If a position of the compared sequences is occupied by the same base or amino acid, the molecules are homologous at that position. The degree of homology between sequences is a function of the number of matching or homologous positions that the sequences share.
[0019]
[0052] "Operably linked" refers to the arrangement of elements, e.g., barcode sequences, gene expression cassettes, coding sequences, promoters, enhancers, transcription factor binding sites, such that the components so described are configured to perform their normal functions. Thus, a control sequence operably linked to a coding sequence can effect transcription, and in some cases translation, of the coding sequence. A control sequence need not be contiguous with the coding sequence, so long as it functions to direct the expression of the coding sequence. Thus, for example, an intervening untranslated but transcribed sequence can be present between the promoter sequence and the coding sequence, and the promoter sequence can still be considered "operably linked" to the coding sequence. Indeed, such sequences need not be present on the same contiguous DNA molecule (i.e., chromosome), and can still have interactions that result in regulatory changes.
[0020]
[0053] The term "selection marker" as used herein refers to a gene introduced into a cell that confers a trait suitable for artificial selection. Commonly used selection markers are well known to those of skill in the art. Drug selection markers such as ampicillin / carbenicillin, kanamycin, chloramphenicol, erythromycin, tetracycline, gentamicin, bleomycin, streptomycin, puromycin, hygromycin, blasticidin, and G418 may be used. Selection markers may also be auxotrophic selectable markers, where the selected cell line carries a mutation that renders it unable to synthesize an essential nutrient. Such strains will only grow if the essential nutrient that they are lacking is provided in the growth medium. For example, auxotrophic selection of yeast mutant strains for essential amino acids is common and well known in the art. "Selection medium" as used herein refers to a cell growth medium to which a compound or biological moiety that selects for or against the selection marker has been added, or a medium that lacks an essential nutrient and selects against auxotrophic strains.
[0021]
[0054] As used herein, the term "vector" refers to any of a variety of nucleic acids containing a desired sequence that is delivered to and / or expressed in a cell. Vectors are typically composed of DNA, although RNA vectors are also available. Vectors include, but are not limited to, plasmids, fosmids, phagemids, viral genomes, BACs, YACs, PACs, synthetic chromosomes, and the like, among others.
[0022]
[0055] As used herein, "affinity" is the strength of the binding interaction between a single biomolecule and its ligand or binding partner. Affinity is usually measured by the equilibrium dissociation constant, K D It is measured and described using K D The lower the value, the higher the affinity between the protein and its binding partner. Affinity can be affected by hydrogen bonds, electrostatic interactions, hydrophobic and van der Waals forces between the binding partners, or by the presence of other molecules such as binding agonists or antagonists.
[0023]
[0056] In some implementations, affinity may be described using arbitrary units, where a particular binding affinity in an assay, such as the binding affinity between two wild-type protein binding partners, or the binding affinity between a wild-type species of a first protein binding partner and a wild-type species of a second protein binding partner, is set to the arbitrary unit of 1.0, and the binding affinity for other pairs of protein binding partners, such as a mutant species of the first protein binding partner and a mutant species of the second protein binding partner, is measured proportionally to that particular binding affinity.
[0024]
[0057] As used herein, "site-saturation mutagenesis" (SSM) refers to a mutagenesis technique used in protein engineering and molecular biology, in which a codon or set of codons is substituted with all possible amino acids at a position in a polypeptide. Alternatively, SSM may represent the change of an amino acid residue at a given position to one of a subset of possible amino acid substitutions at that position, e.g., substitution with all possible amino acids except cysteine. SSM may be performed on one codon, several codons, or all positions in a protein. The result is a library of mutant proteins representing the full complement of possible amino acids at one, several, or all amino acid positions in a polypeptide.
[0025]
[0058] As used herein, "user-directed mutagenesis" refers to any process in which a user modifies the amino acid sequence of a polypeptide by any technique known to one of skill in the art. A polypeptide can be modified at one or more amino acid residues in a defined manner, e.g., an alanine residue is changed to an arginine residue, or the polypeptide sequence can be modified in a randomized manner, e.g., using degenerate primers and randomized PCR amplification. A polypeptide can be modified by user-directed mutagenesis at one amino acid residue or many amino acid residues. A polypeptide can be modified by user-directed mutagenesis to include an insertion and / or deletion of one or more amino acid residues, or a polypeptide sequence can be truncated by user-directed mutagenesis. A polypeptide can be modified by user-directed mutagenesis to include an insertion or substitution of a natural or non-natural amino acid.
[0026]
[0059] As used herein, a "paratope" is a portion of an antibody that specifically recognizes and binds to the antibody's corresponding antigen. A paratope is also known as an antigen-binding site. A paratope may comprise as many as about 15 amino acid residues of an antibody polypeptide, of which about 5 amino acid residues usually contribute the majority of the binding energy to the paratope. The amino acids that comprise the paratope may be a contiguous sequence of amino acid residues within the polypeptide chain of the antibody protein structure, or may be discontinuous amino acid residues that confer conformational specificity to the three-dimensional structure of the antibody protein structure. As used herein, "paratope mapping" is the process of experimentally identifying and characterizing the composition of a paratope within an antibody protein structure. Paratope mapping may define the amino acid sequence of the paratope, the three-dimensional structure of the paratope, and may provide mechanistic information that defines the interaction of an antibody with its antigen.
[0027]
[0060] As used herein, an "epitope" is a portion of an antigen that is specifically recognized and bound by an antibody. An epitope may comprise as many as about 15 amino acid residues of an antigenic polypeptide, of which about 5 amino acid residues usually provide the majority of the binding energy to the epitope. The amino acids that comprise an epitope may be a contiguous sequence of amino acid residues within the polypeptide chain of an antigenic protein, or may be discontinuous amino acid residues that provide conformational specificity to the three-dimensional structure of the folded antigenic protein.
[0028]
[0061] As used herein, "epitope mapping" is the process of experimentally identifying and characterizing the composition of epitopes within an antigenic protein. Epitope mapping can define the amino acid sequence of the epitope, the three-dimensional structure of the epitope, and can provide mechanistic information that defines the interaction of an antibody with its antigen.
[0029]
[0062] As used herein, a "receptor" is a chemical structure that includes a polypeptide sequence that receives and transmits signals relevant to a biological system in its natural physiological context. Receptors are a diverse class of proteins and can include transmembrane receptors, intracellular receptors, cytoplasmic receptors, nuclear receptors, and the like. Transmembrane receptors are located in the plasma membrane, with a portion of the receptor located outside the cell to receive signals from outside the cell. Receptors receive and transmit signals through a wide variety of mechanisms, including, but not limited to, signals transmitted by ligand-gated ion channels, G protein-coupled receptors, kinase-associated receptors, or by translocation of the receptor across the nuclear envelope. Receptors usually bind to specific ligands, which can be agonists, partial agonists, antagonists, inverse agonists, or corresponding allosteric modulators of the receptor.
[0030]
[0063] As used herein, a "ligand" is a molecule that produces a signal by binding to a receptor. A ligand molecule can be a polypeptide, an inorganic molecule, or an organic molecule. In some cases, ligand binding to a receptor protein changes the conformation of the protein to produce and transmit a signal across or within the cell. Ligands can include substrates, inhibitors, activators, signaling lipids, neurotransmitters, and other molecules. In many cases, the binding of a ligand to its corresponding receptor is specific with high binding affinity.
[0031]
[0064] As used herein, a "wild-type protein binding partner" is one of two polypeptides that specifically interact with each other in a biological context. As used herein, a "wild-type protein binding interaction" is an interaction between two wild-type protein binding partners. Wild-type protein binding partners include full-length human proteins; full-length proteins of other animal species; truncated proteins of any animal species; parts of proteins of any animal species; plant proteins, fungal proteins, viral proteins, de novo proteins, or truncated species of proteins of any origin. A wild-type protein binding partner can be a synthetic peptide, a glycosylated polypeptide, or a polypeptide with other synthetic or natural post-translational modifications. A wild-type protein binding partner can be part of an engineered polypeptide, such as an antibody that has been engineered to produce a therapeutic effect. As used herein, a wild-type protein binding partner can include naturally occurring variations of an animal's polypeptide sequence, such as natural variants due to SNPs or indels in the coding nucleotide sequence.
[0032]
[0065] As used herein, a "mutant protein binding partner" is one of two modified polypeptides whose unmodified species specifically interact with each other in a biological context. One or both protein binding partners in a wild-type protein binding interaction can be modified to produce a mutant protein binding partner. A mutant protein binding partner may or may not interact with the corresponding wild-type species of that protein binding partner. A mutant protein binding partner may or may not interact when both protein binding partners of a wild-type protein binding interaction are modified to produce a first mutant protein binding partner and a second mutant protein binding partner. A wild-type protein binding partner can be modified by user-directed mutagenesis or site-saturation mutagenesis to produce a mutant protein binding partner.
[0033]
[0066] In some implementations, the method includes a first protein binding partner and a library of second protein binding partners. The library of second protein binding partners includes a plurality of user-specified or randomly added variants of a protein and a wild-type protein. The plurality of user-specified or randomly added variants of a protein can include variants of the protein having 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, or more amino acid substitutions. The amino acid substitutions can be selected to introduce a charge change to the protein or a conformational change to the protein, and the wild-type amino acid can be replaced with a natural or unnatural amino acid. In some implementations, the amino acid substitutions can be generated by site saturation mutagenesis (SSM) to create the SSM library of protein binding partners. In some implementations, the library of second protein binding partners can be created by alanine scanning. In some implementations, the library of second protein binding partners can be created by random mutagenesis, such as using error-prone PCR, or by other methods to introduce mutations into the amino acid sequence of the expressed protein. The first protein binding partner and the library of second protein binding partners are assayed for binding affinity in a parallelized high-throughput manner such that the affinity is measured individually for the interaction between the first protein binding partner and each of the plurality of user-specified mutants. Members of the library of second protein binding partners that are found to have a binding affinity with the first protein binding partner that is higher or lower than the binding affinity of the first protein binding partner to the wild-type target protein are identified and selected for further study.
[0034]
[0067] In some implementations where a library of first and second protein binding partners are assayed for binding affinity, the assay can be phage display, yeast surface display, or another parallelized high-throughput method.
[0035]
[0068] In other implementations, the method includes a library of first protein binding partners and a library of second protein binding partners. The library of first protein binding partners includes a plurality of user-specified or randomly added variants of a protein and a wild-type protein. The plurality of user-specified or randomly added variants of a protein can include variants of a targeted protein having 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, or more amino acid substitutions. The amino acid substitutions can be selected to introduce a charge change to the protein or a conformational change to the protein, and the wild-type amino acid can be replaced with a natural or unnatural amino acid. In some implementations, the amino acid substitutions can be generated by site saturation mutagenesis (SSM), and the SSM library of protein binding partners can be created. In some implementations, the library of second protein binding partners can be created by alanine scanning. In some implementations, the library of second protein binding partners can include a plurality of user-specified or randomly added variants of a protein and a wild-type protein. The plurality of user-specified or randomly added variants of a protein may include variants of a target protein having one, two, three, four, five, six, seven, eight, nine, ten, or more amino acid substitutions. The amino acid substitutions may be selected to introduce a charge change or a conformational change to the protein, and a wild-type amino acid may be replaced with a natural or non-natural amino acid. In some implementations, the amino acid substitutions may be generated by site-saturation mutagenesis (SSM), and an SSM library of protein-binding partners may be created. In some implementations, a library of second protein-binding partners may be created by alanine scanning. The library of first protein-binding partners and the library of second protein-binding partners are assayed for binding affinity in a parallelized high-throughput method, such that the affinity is measured individually and pairwise for the interaction between each of the plurality of mutant first protein-binding partners and each of the plurality of mutant second protein-binding partners.Pairs comprising a member selected from the library of a first protein binding partner and a member selected from the library of a second protein binding partner that are found to have a binding affinity that is higher or lower than the binding affinity of the wild-type first protein binding partner and the wild-type second protein binding partner are identified and selected for further study.
[0036]
[0069] In some implementations where a library of first protein binding partners is assayed against a library of second protein binding partners for binding affinity, the assay can be by the yeast two-hybrid system, the AlphaSeq system, or another parallelized high-throughput library-by-library screening method. Binding affinity for interactions between mutant protein binding partners versus wild-type protein binding partners can be measured by any number of methods to quantify protein binding affinity, such as yeast two-hybrid screening, biolayer interferometry, ELISA, quantitative ELISA, surface plasmon resonance, FACS-based enrichment methods, synthetic yeast agglutination reactions, the AlphaSeq platform, or any other measurement of protein interaction strength. The AlphaSeq method is described in U.S. Patent Application Serial No. 15 / 407,215 (U.S. Patent Application Publication No. 2017-0205421A1), which is incorporated herein in its entirety for all purposes.
[0037]
[0070] In some implementations, protein-binding partner pairs identified by the methods disclosed herein are further characterized, for example, by crystallography, cryo-electron microscopy, microelectron diffraction, mass spectrometry, computer modeling, among other methods for characterizing protein-protein complexes known in the art. Protein-binding partner or mutant protein-binding partner pairs can be further characterized individually or in the context of a protein-protein complex between the two partners.
[0038]
[0071] In some implementations, the first binding partner and the second protein binding partner are full-length proteins. In other implementations, the first binding partner and the second protein binding partner are truncated proteins. In other implementations, the first binding partner and the second protein binding partner are fusion proteins. In other implementations, the first binding partner and the second protein binding partner are tagged proteins. Tagged proteins include epitope-tagged, e.g., FLAG-tagged, HA-tagged, His-tagged, Myc-tagged proteins, among others known in the art. In some implementations, the first protein binding partner is a full-length protein and the second protein binding partner is a truncated protein. Each of the first protein binding partner and the second protein binding partner can be a full-length protein, a truncated protein, a fusion protein, a tagged protein, or a combination thereof.
[0039]
[0072] In some implementations, the first binding partner is an antibody or a truncated portion of an antibody polypeptide. In other embodiments, the library of first binding partners is an antibody, a library of truncated antibody polypeptides, or a library of antibody variants created by site saturation mutagenesis, alanine scanning, or other methods well known in the art. Antibodies, also known as immunoglobulins, are relatively large multi-unit protein structures that specifically recognize and bind to a unique molecule or molecules. For most antibodies, two heavy chain polypeptides of approximately 50 kDA and two light chain polypeptides of approximately 25 kDA are linked by disulfide bonds to form a larger Y-shaped multi-unit structure. The variable and hypervariable regions, which represent the variability of amino acid sequences at the tips of the Y-shaped structure, confer specificity for a given antibody to recognize its target.
[0040]
[0073] In some implementations, the first binding partner is a single chain variable fragment (scFv), a fusion protein of the variable regions of the heavy (VH) and light (VL) chains of an immunoglobulin linked by a short linker peptide. In some implementations, the library of first protein binding partners is a library of scFvs, or a library of scFv variants generated by site-saturation mutagenesis, alanine scanning, or other methods well known in the art.
[0041]
[0074] In some implementations, the first binding partner is an antigen-binding fragment (Fab), the region of an antibody that binds to the antigen. A Fab can contain one constant domain and one variable domain each of the heavy and light chains, and contains the paratope region of the antibody. In some implementations, the library of first protein binding partners is a library of Fabs, or a library of Fab variants created by site-saturation mutagenesis, alanine scanning, or other methods well known in the art.
[0042]
[0075] In some implementations, the first binding partner can be a portion of a single domain antibody, or a VHH, an antigen-binding fragment of an antibody that is only a heavy chain. A VHH comprises one variable domain of a heavy chain antibody. In some implementations, the library of first protein binding partners is a library of VHHs, or a library of VHH variants created by site-saturation mutagenesis, alanine scanning, or other methods well known in the art.
[0043]
[0076] In some implementations, the second binding partner is an antigen. In other implementations, the library of second binding partners is a library of antigens, or a library of antigens created by site-saturation mutagenesis, among other methods. An antigen is a molecule or molecular structure that is targeted by an antibody. An antigen is usually a protein, polypeptide, or polysaccharide that is targeted by a corresponding specific antibody. An antigen includes an epitope, the portion of an antigen that is recognized by the corresponding antibody and confers specificity to the antigen.
[0044]
[0077] In some implementations, for a pair of protein binding partners where the first protein binding partner is an antibody, scFv, Fab, or FHH and the second protein binding partner is an antigen, a wild-type antibody scFv, Fab, or FHH, the wild-type antibody scFv, Fab, or FHH may be screened against a library of mutant antigens to determine the effect of the antigen mutants on the affinity of the antibody to the antigen. In other implementations, the wild-type antibody, scFv, Fab, or FHH may be screened against a library of mutant antigens for the purpose of epitope mapping, i.e., the amino acid sequence of the epitope, the three-dimensional structure of the epitope may be defined, providing information regarding the mechanism of action that defines the interaction of the epitope with the antibody.
[0045]
[0078] In some implementations, for a pair of protein binding partners where the first protein binding partner is an antibody, scFv, Fab, or FHH and the second protein binding partner is an antigen, a library of mutated antibodies, the scFv, Fab, or FHH may be screened against a library of wild-type antigen to determine the effect of the antibody, scFv, Fab, or FHH mutants on the affinity of the antibody, scFv, Fab, or FHH to the antigen. In other implementations, a library of mutated antibodies, scFv, Fab, or FHH may be screened against the wild-type antigen for the purpose of paratope mapping, i.e., the amino acid sequence of the paratope, the three-dimensional structure of the paratope may be defined, providing information regarding the mechanism of action that defines the interaction of the paratope with the antigen.
[0046]
[0079] In some implementations, for a pair of protein binding partners where the first protein binding partner is an antibody, scFv, Fab, or FHH and the second protein binding partner is an antigen, a library of mutated antibodies, the scFv, Fab, or FHH may be screened against a library of mutated antigens to simultaneously examine the effect of the antibody, scFv, Fab, or FHH mutants on the affinity of the antibody, scFv, Fab, or FHH to the antigen. In other implementations, a library of mutated antibodies, scFv, Fab, or FHH may be screened against a library of mutated antigens for the purposes of epitope and paratope mapping, i.e., the amino acid sequences of the epitopes and paratopes, the three-dimensional structures of the epitopes and paratopes may be defined, providing information regarding the mechanism of action that defines the interaction of the antibody with the antigen.
[0047]
[0080] As used herein, "substantially different from" refers to two quantitative binding affinity values that differ from each other by about 5%, 10%, 20%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 60%, 70%, 80%, 90%, 100%, 200%, 300% to about 500% or more in magnitude. D It can be measured in arbitrary units or quantified by normalizing the binding affinity of a particular pair of protein-binding partners to 1.0 arbitrary units and measuring the binding affinities of multiple other protein-binding partners in arbitrary units to that particular pair of protein-binding partners normalized to 1.0 arbitrary units.
[0048]
[0081] As used herein, "substantially the same as" refers to two quantitative binding affinity values that are within about 20%, 15%, 10%, 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2%, 1% or about 0.1% of the value. DIt can be measured in arbitrary units or quantified by normalizing the binding affinity of a particular pair of protein-binding partners to 1.0 arbitrary units and measuring the binding affinities of multiple other protein-binding partners in arbitrary units to that particular pair of protein-binding partners normalized to 1.0 arbitrary units.
[0049]
[0082] As used herein, "substantially higher" refers to a quantitative binding affinity value that is about 5%, 10%, 20%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 60%, 70%, 80%, 90%, 100%, 200%, 300% to about 500% or more higher than one quantitative binding affinity value. D It can be measured in arbitrary units or quantified by normalizing the binding affinity of a particular pair of protein-binding partners to 1.0 arbitrary units and measuring the binding affinities of multiple other protein-binding partners in arbitrary units to that particular pair of protein-binding partners normalized to 1.0 arbitrary units.
[0050]
[0083] As used herein, "substantially lower than" refers to a quantitative binding affinity value of one that is about 95%, 90%, 85%, 80%, 75%, 70%, 65%, 60%, 55%, 50%, 45%, 40%, 35%, 30%, 25%, 20%, 15%, 10% to about 5% or less of the quantitative binding affinity value of the other. D It can be measured in arbitrary units or quantified by normalizing the binding affinity of a particular pair of protein-binding partners to 1.0 arbitrary units and measuring the binding affinities of multiple other protein-binding partners in arbitrary units to that particular pair of protein-binding partners normalized to 1.0 arbitrary units.
[0051]
[0084] In some implementations, the methods disclosed herein can be used to identify compensatory mutations in protein binding partners. As described above, the methods disclosed herein can be used to screen a library of first protein binding partners against a library of second protein binding partners such that the affinity of the interaction between each of the plurality of first protein binding partners and each of the second protein binding partners is measured in a parallelized high-throughput manner. For a given interaction between two individual species of protein binding partners, the following affinity relationships may occur simultaneously: (a) a variant of the first protein binding partner and a wild-type species of the second protein binding partner have a binding affinity, as detected by the methods disclosed herein, that is lower than the binding affinity between the wild-type species of the first protein binding partner and the wild-type species of the second protein binding partner; (b) a wild-type species of the first protein binding partner and a wild-type species of the second protein binding partner have a binding affinity, as detected by the methods disclosed herein, that is lower than the binding affinity between the wild-type species of the first protein binding partner and the wild-type species of the second protein binding partner; and (c) a variant of the first protein binding partner as described in (a) and a variant of the second protein binding partner as described in (b) have a binding affinity, as detected by the methods disclosed herein, that is stronger than, equal to, or nearly equal to the binding affinity between the wild-type species of the first protein binding partner and the wild-type species of the second protein binding partner. Two mutations in a pair of protein-binding partners that exhibit the above-mentioned relationship may be referred to as compensatory mutations, such that when the two mutations occur together, the mutation in the second protein-binding partner compensates for the affinity-reducing effect of the mutation in the first protein-binding partner, thereby restoring wild-type affinity levels between the two protein-binding partners, as illustrated in Figure 15. This scenario implies proximity between the two mutated residues and may be useful for structure determination and / or protein engineering.
[0052]
[0085] In another implementation, for a given interaction between two respective species of protein binding partners, the following alternative affinity associations may be simultaneously detected: (a) a variant of the first protein binding partner and a wild-type species of the second protein binding partner have a binding affinity, as detected by the methods disclosed herein, that is lower than the binding affinity between the wild-type species of the first protein binding partner and the wild-type species of the second protein binding partner; (b) a wild-type species of the first protein binding partner and a variant of the second protein binding partner have a binding affinity, as detected by the methods disclosed herein, that is stronger than, equal to, or approximately equal to the binding affinity between the wild-type species of the first protein binding partner and the wild-type species of the second protein binding partner; and (c) a variant of the first protein binding partner as described in (a) and a variant of the second protein binding partner as described in (b) have a binding affinity, as detected by the methods disclosed herein, that is stronger or significantly stronger than the binding affinity between the wild-type species of the first protein binding partner and the wild-type species of the second protein binding partner. Two mutations in a pair of protein-binding partners that exhibit the above-mentioned relationship may be called compensatory mutations, where the mutations in the protein-binding partners together confer additional binding affinity that is higher than either of the two compensatory mutations occurring by themselves. This scenario is shown in Figure 16 between the K54I mutation in the antigen PD-1 and the Y101K mutation in the short chain variable fragment (scFv) of the monoclonal antibody pembrolizumab (Pembro). This scenario implies the proximity between the two mutated residues, which may be useful for structural determination, protein engineering, or IP protection purposes.
[0053]
[0086] In another implementation, for a given interaction between two individual species of protein binding partners, the following alternative affinity relationships may be simultaneously detected: (a) the variant of the first protein binding partner and the wild-type species of the second protein binding partner have a binding affinity that is stronger than, equal to, or nearly equal to the binding affinity of the wild-type species of the first protein binding partner and the wild-type species of the second protein binding partner; (b) the wild-type species of the first protein binding partner and the variant of the second protein binding partner have a binding affinity that is lower than the binding affinity of the wild-type species of the first protein binding partner and the wild-type species of the second protein binding partner; and (c) the variant of the first protein binding partner described in (a) and the variant of the second protein binding partner described in (b) have a binding affinity detected by the methods disclosed herein that is equal to, or nearly equal to, the binding affinity of the wild-type species of the first protein binding partner and the wild-type species of the second protein binding partner. Two mutations in a pair of protein-binding partners that exhibit the above-mentioned relationship may also be referred to as compensatory mutations, such that when the two mutations occur simultaneously, the mutation in the second protein-binding partner compensates for the affinity-reducing effect of the mutation in the first protein-binding partner, thereby restoring the wild-type affinity level between the two protein-binding partners.
[0054]
[0087] In another implementation, for a given interaction between two individual species of protein binding partners, the following alternative affinity relationships may be simultaneously detected: (a) the variant of the first protein binding partner and the wild-type species of the second protein binding partner have a binding affinity that is stronger than, equal to, or nearly equal to the binding affinity of the wild-type species of the first protein binding partner and the wild-type species of the second protein binding partner; (b) the wild-type species of the first protein binding partner and the variant of the second protein binding partner have a binding affinity that is stronger than, equal to, or nearly equal to the binding affinity of the wild-type species of the first protein binding partner and the wild-type species of the second protein binding partner; and (c) the variant of the first protein binding partner described in (a) and the variant of the second protein binding partner described in (b) have a binding affinity, as detected by the methods disclosed herein, that is lower than the binding affinity of the wild-type species of the first protein binding partner and the wild-type species of the second protein binding partner. Two mutations in a pair of protein-binding partners that exhibit the above-mentioned relationships can be useful for identifying amino acids that are in close proximity to each other at the interface between the proteins, and are particularly useful in mediating binding affinity between the two protein-binding partners.
[0055]
[0088] In some implementations, an "expected binding affinity" or "expected interaction strength" may be defined and predicted for a pair of mutated protein-binding partners. In some implementations, an expected binding affinity may be defined for the pairing of an antibody variant and an antigen variant. As used herein, an expected binding affinity may be defined as the affinity that would be expected to be observed between two mutated protein-binding partners based on the observed effect of each mutation on binding to the corresponding wild-type protein-binding partner. The expected binding affinity is calculated by (1) normalizing the binding affinity for each wild-type to 1.0, (2) calculating the relative binding affinity of each of the interactions of the mutated protein-binding species with its wild-type protein-binding partner to obtain a first mutated protein binding affinity and a second mutated protein binding affinity, and (3) multiplying the first mutated protein binding affinity and the second mutated protein binding affinity to obtain the expected binding affinity for the interaction of the two protein-binding partners with each other.
[0056]
[0089] For example, the binding affinity observed for the interaction between a wild-type species of a first protein binding partner and a wild-type species of a second protein binding partner is normalized to an arbitrary unit of 1.0; the binding affinity observed for the interaction between a wild-type species of a first protein binding partner and a mutant species of a second protein binding partner is 0.5 relative to the affinity of the wild-type protein binding interaction; the binding affinity observed for the interaction between a mutant species of a first protein binding partner and a wild-type species of a second protein binding partner is 0.5 relative to the affinity of the wild-type protein binding interaction; and the binding affinity observed for the interaction between a mutant species of a first protein binding partner and a mutant species of a second protein binding partner is calculated to be 0.25.
[0057]
[0090] In some implementations, a "measured binding affinity" can be determined for each of many interactions between mutant protein binding partners according to the methods disclosed herein. In one implementation, the affinity measurement of the interaction between a wild-type species of a first protein binding partner and a wild-type species of a second protein binding partner is normalized to an arbitrary unit of 1.0. The measured binding affinities of other pairs of protein binding partners, such as the binding affinities of a mutant species of a first protein binding partner and a mutant species of a second protein binding partner, are measured and quantified proportionally to the value of 1.0 assigned to the interaction between a wild-type species of a first protein binding partner and a wild-type species of a second protein binding partner. The measured binding affinities of pairs of mutant protein binding partners can be compared to the expected binding affinities to determine the ratio of the measured binding affinity to the expected binding affinity. In some implementations, and for some pairs of protein-binding partners, the ratio of measured binding affinity to expected binding affinity can be about 1:1, 2:1, 3:1, 4:1, 5:1, 6:1, 7:1, 8:1, 9:1 to about 10:1, or greater than 10:1. In some implementations, and for some pairs of protein-binding partners, the ratio of measured binding affinity to expected binding affinity can be about 1:1, 2:1, 3:1, 4:1, 5:1, 6:1, 7:1, 8:1, 9:1 to about 10:1, or less than 10:1.
[0058]
[0091] For protein binding partner pairs where the first protein binding partner is an antibody and the second protein binding partner is an antigen, and where compensatory mutations of the antibody and antigen have been identified by the methods disclosed herein, the amino acid residues involved in these pairs of compensatory mutations are spatially close at the antigen / antibody interface, providing unique information about the protein-protein interface that is not available using methods based on one-sided protein binding. Examples of compensatory mutations between protein binding partners detected by the methods disclosed herein are indicative of structural proximity. In the absence of other structural data, pairs of compensatory mutations can be useful as distance constraints in building computational models of protein-protein interactions. Identifying compensatory mutations of pairs of protein binding partners provides unique information about the proximity of interacting residues at the protein-protein interface. These distance constraints can also be useful for protein engineering and structure determination, or to inform intellectual property protection efforts for novel antibodies or antigens in the pharmaceutical and biotechnology industries.
[0059]
[0092] In some implementations, the methods disclosed herein can be used to identify compensatory mutations between protein binding partners, where the first protein binding partner is a receptor and the second protein binding partner is a ligand. The amino acid residues involved in these pairs of compensatory mutations are spatially close at the receptor / ligand interface, providing unique information about the protein-protein interface that is not available when using methods based on one-sided protein binding. Examples of compensatory mutations between protein binding partners detected by the methods disclosed herein are indicative of structural proximity. In the absence of other structural data, pairs of compensatory mutations can be useful as distance constraints in building computational models of protein-protein interactions. Identifying compensatory mutations of pairs of protein binding partners provides unique information about the proximity of interacting residues at the protein-protein interface. These distance constraints can also be useful for protein engineering, structure determination, or to inform rational design efforts for novel receptors and ligands in the pharmaceutical and biotechnology industries. The compensatory mutations identified for the receptor-ligand protein binding partners can be used to custom engineer specific behaviors between receptor-ligand interactions that are useful for biomedical applications, such as cell therapy, cancer therapy, immunotherapy, etc. In some implementations, compensatory mutations can be identified between receptor-ligand protein binding partners, where the receptor-ligand protein binding partner containing the compensatory mutation exhibits higher affinity than between the wild-type species of the protein binding partners.
[0060]
[0093] The methods disclosed herein are uniquely advantageous for identifying such synergistic interactions, i.e., mutations that increase the binding affinity between two protein binding partners, e.g., a receptor and its ligand. Using previously available methods, e.g., traditional one-sided screening methods, it has been very difficult or impossible to identify such synergistic compensatory mutations between protein binding partners.
[0061]
[0094] Additionally, the methods disclosed herein may be useful for identifying and engineering orthogonal protein interactions, for example between a cell surface receptor and its ligand, such that the interaction of the engineered receptor, engineered ligand, and endogenous wild-type ligand (e.g., a soluble growth factor or cytokine) can be uniquely tailored to a desired outcome in a therapeutic context. For example, the protein interaction illustrated by FIG. 3B represents a one-sided orthogonal binding relationship, where the wild-type receptor binds and is activated by the wild-type ligand, but not the mutant ligand, whereas the mutant ligand binds and is activated by both the wild-type and mutant ligands. The methods disclosed herein allow for the identification of mutations in both the receptor and the ligand, and a small number of highly impactful mutations may be introduced into the receptor and ligand to confer such properties on the receptor-ligand interaction.
[0062]
[0095] The one-sided orthogonal binding relationship illustrated by FIG. 3B may be particularly useful in the context of cell therapy, e.g., CAR-T cell therapy, where regulating the number and abundance of CAR-T cells in a patient may be important to the efficacy of the therapy. Using the methods disclosed herein, compensatory mutations for the receptor may be identified, allowing the engineering of CAR-T cells that express customized cell surface receptor-associated compensatory mutations identified by the methods disclosed herein. Similarly, soluble growth factors or cytokines may be engineered to express compensatory mutations identified by the methods disclosed herein, such that the CAR-T cell surface receptor and the soluble growth factor or cytokine exhibit a one-sided one-sided orthogonal affinity relationship as shown in FIG. 3B. By introducing a small number of potentially highly impactful compensatory mutations into each of the cell surface receptor and the soluble growth factor or cytokine, the CAR-T cell surface receptor may bind and be activated by both the engineered growth factor or cytokine, and the native wild-type growth factor or cytokine to the patient's physiological environment. Conversely, engineered soluble growth factor or cytokine-associated compensatory mutations identified by the methods disclosed herein bind and activate only the engineered CAR-T cell surface receptor, without affecting multiple or wild-type cell surface receptors on the patient's physiology. This pattern of customized orthogonal protein-protein interactions using the methods disclosed herein will be useful in engineering cell therapies, immunotherapies, and biologics to treat a number of diseases and disorders.
[0063]
[0096] FIG. 1 is a series of charts showing the library-by-library screening capabilities of the AlphaSeq method. In each chart, a subset of protein interactions with affinities measured by BioLayer Interferometry across a wide affinity range are compared to AlphaSeq Intensity to show the sensitivity and quantitative accuracy of the AlphaSeq method at a given meshwork size. Chart 100 illustrates screening the interaction of a first library of 100 binding partners against a second library of 100 binding partners and measuring 10,000 interactions. Chart 102 illustrates screening the interaction of a first library of 1,000 binding partners against a second library of 1,000 binding partners and measuring 1,000,000 interactions. Chart 104 illustrates screening the interaction of a first library of 10,000 binding partners against a second library of 10,000 binding partners and measuring 100,000,000 interactions. Chart 106 shows the protein-protein affinity (K D ) versus AlphaSeq intensity. Chart 108 shows the protein-protein affinity (K D Chart 110 shows the correlation between protein-protein affinity (K ) and AlphaSeq intensity for 100,000,000 interactions. D ) and AlphaSeq intensity.
[0064]
[0097] 2A is a schematic diagram of two interacting protein binding partners in a complex, where a first protein binding partner 200 is an antibody and a second protein binding partner 204 is an antigen, highlighting the interface between the protein binding partners and a site saturation mutagenesis (SSM) screen of the two protein binding partners 200 and 204. Amino acid residue 202 of protein binding partner 200 corresponds to amino acid residue 203 of protein binding partner 204. Amino acid residue 202 of protein binding partner 200 can be replaced by any available additional amino acid residue, natural or artificial, and screened for interaction against a similar library of replacements of amino acid residue 203 of protein binding partner 204.
[0065]
[0098] The results of such a per-library SSM screen are shown in FIG. 2B. Heat map 206 shows per-library AlphaSeq intensity measurements of interactions between protein-binding partners carrying SSM mutations at all amino acid residues that define the protein-protein interface. Darker colors represent higher AlphaSeq intensity, whereas lighter colors represent lower AlphaSeq intensity. For example, inset 208 highlights the per-library AlphaSeq intensity of an SSM library with a substitution at amino acid 210 measured against an SSM library with a substitution at amino acid 212. For the per-library screen whose data is represented by heat map 206, amino acid residue 210 has been mutated to any of the available natural amino acid residues (G, A, V, L, M, I, S, TC, P, N, Q, F, Y, W, K, R, H, D, E). Corresponding to amino acid residue 210, amino acid residue 212 is similarly mutated to any of the available natural amino acid residues (G, A, V, L, M, I, S, TC, P, N, Q, F, Y, W, K, R, H, D, E). Pairwise interaction strength data for variants of amino acid residue 210 and amino acid residue 212 is represented by heatmap inset 208. For example, a color version of the heatmap included in FIG. 2B is available via the United States Patent and Trademark Office (USPTO) Published Patent Application Information Retrieval System (PAIR, accessible at the following link: https: / / portal.uspto.gov / pair / PublicPair and U.S. Patent Application No. 63 / 033,176, Supplemental Content tab).
[0066]
[0099] 3A-3C are graphical representations of a subset of protein-protein interactions detected by the data shown in FIGS. 2A-2B, illustrating the ability of the methods disclosed herein to detect the relative affinities between wild-type and mutant protein binding partners, and the effect of a single amino acid substitution on the affinity between two protein binding partners. FIG. 3A illustrates a scenario in which the wild-type protein binding partner interacts with high affinity, and the mutant protein binding partner interacts with high affinity, but the mutant of either the first or second protein binding partner does not interact with the wild-type of the other protein binding partner. The result is a pair of mutants, each with a single amino acid change from the wild-type, that bind orthogonally to the wild-type. FIG. 3B illustrates a scenario in which both the wild-type and mutant forms of the first protein binding partner interact with the wild-type of the second protein binding partner, but the wild-type first protein binding partner does not interact with the mutant second protein binding partner, i.e., the mutation of the second protein binding partner abolishes the interaction of the wild-type first protein binding partner. FIG. 3C illustrates a scenario in which both the wild-type and mutant forms of a first protein binding partner interact with a mutant form of a second protein binding partner, but the mutant form of the first protein binding partner does not interact with the wild-type second protein binding partner, i.e., the mutation of the first protein binding partner abolishes interaction with the wild-type second protein binding partner.
[0067]
[0100] Figure 4 shows the workflow of library-by-library protein-protein interaction screening using AlphaSeq. A first library of protein-binding partners 400 and a second library of protein-binding partners 402 are generated by site-saturation mutagenesis and expressed in yeast. In an interaction step 404, the two library populations are mixed and the protein-binding partners bind. Cells expressing interacted protein-binding partners are mated in a fusion step 406. Protein-protein interactions between the first and second libraries are detected and quantified in a measurement step 408.
[0068]
[0101] FIG. 5 illustrates that antibody-antigen interactions can be measured with the AlphaSeq platform. Well-characterized antibody-antigen pairs well known in the art were subjected to the AlphaSeq workflow. The system identified pairs of cells with cognate binding partners and did not detect cross-reactivity among non-cognate pairs. Plot 500 shows the detected interaction of huCTLA-4 and ipilimumab scFv, as well as the relative AlphaSeq signal. Plot 502 shows the detected interaction of huTNFα and adalimumab scF.
[0069]
[0102] Figure 6 illustrates the results of an AlphaSeq experiment screening eight antigen variants against eight antibody variants, and 64 interactions were detected and quantified. The thickness of the connecting lines indicates the magnitude of the mating frequency signal. Significant line 600 is a modified version of Maute et al. (Maute RL, Gordon SR, Mayer AT, McCracken MN, Natarajan A, Ring NG, Kimura R, Tsai JM, Manglik A, Kruse AC, Gambhir SS, Weissman IL, Ring AM. Engineering high-affinity PD-1 variants for optimized immunotherapy and immuno-PET imaging. Proc Natl Acad Sci US A. 2015 Nov 24;112(47):E6506-14. doi: 10.1073 / pnas.1519623112. Epub 2015 Nov 10. PMID: 26604307; PMCID: Figure 3 represents the interaction between human programmed cell death ligand-1 (huPD-L1) and engineered programmed cell death protein-1 (PD-1) ectodomain, as previously reported and characterized by huPD-L1 (PMC4664306.) The significance of this interaction, as detected by the AlphaSeq platform, confirms that the methods disclosed herein can detect interactions between protein binding partners, which are enhanced relative to the wild-type interaction by modification of one or both of the protein binding partners.
[0070]
[0103] Figure 7 is a heat map showing the results of a screen of 60 PD-1 variants (antigen variants) against wild-type pembro scFv (antibody). 60 PD-1 surface residues were selected for mutagenesis and the resulting SSM library was subjected to the AlphaSeq workflow. The AlphaSeq signal is displayed in a heat map format, with dark shaded boxes of various patterns indicating high and low joining frequencies. The bars at the bottom of the figure represent the shortest distance between the residue and any atom in the antibody. Residues 700 and 702, which correspond to PD-1 residues 54 and 61, are particularly intolerant to substitutions and are spatially close to the antibody based on the known X-ray structure. For example, a color version of the heat map included in Figure 7 is available via USPTO PAIR (accession 63 / 033,176, additional content tab).
[0071]
[0104] Figure 8 highlights specific residues in the crystal structure of the PD-1 / pembrolizumab interface that were identified by the data presented in Figure 7. PD-1 residues K54 and D61 are particularly intolerant to substitutions and are spatially close in the antibody-antigen interface.
[0072]
[0105] FIG. 9 is a heat map showing the results of a screen of 33 pembrolizumab scFv variants (antibody variants) against wild-type PD-1 (antigen). Thirty-three positions in pembrolizumab scFv were mutagenized using SSM and the resulting library was subjected to the AlphaSeq workflow against wild-type PD-1. The AlphaSeq signal is displayed in a heat map format, with various patterns of dark shaded boxes indicating high and low joining frequencies. The bars at the bottom of the figure represent the shortest distance between the residue and any atom in the antibody. Pembrolizumab scFv residue 99 is particularly intolerant to substitutions and is spatially close in the antibody-antigen interface, as indicated by the shaded column 900 and shaded box 902. A color version of the heat map included in FIG. 9 is available via USPTO PAIR (accession 63 / 033,176, additional content tab).
[0073]
[0106] Figure 10 is a graphical representation of the crystal structure of the PD-1 / pembrolizumab scFv interface, highlighting specific residues at the antibody-antigen interface. PD-1D61 and pembrolizumab scFvR99 have been shown to be functionally important for productive antigen-antibody complex formation, and in certain embodiments, substitutions at either site significantly reduce mating frequency in AlphaSeq assays. For example, a color version of the heatmap included in Figure 10 is available via USPTO PAIR (Accession 63 / 033,176, Additional Content tab).
[0074]
[0107] Figure 11 is a structural illustration of the PD-1 / pembrolizumab interface highlighting the dense network of interactions surrounding the D61-R99 pair of residues highlighted above. Mutation-intolerant residues of PD-1 and pembrolizumab scFv are shown.
[0075]
[0108] Figures 12A-12B are graphical depictions of the same data set shown in Figure 9. Pembrolizumab scFv residues D104 and S230 are highlighted as being particularly intolerant to amino acid substitutions. These residues interact with each other across the VH-VL interface, forming interactions that stabilize the relative positioning of the VH and VL domains within the antibody structure. Disruption of this specific interaction by mutation results in loss of binding, as read by a decrease in the joining frequency score generated by AlphaSeq. It is noteworthy that while alanine substitution at pembrolizumab scFv residue 230 is tolerated, most other substitutions are not. Alanine scanning alone does not identify this site as amenable to mutation, highlighting the advantage of using site saturation mutagenesis and the full mutation spectrum generated by the AlphaSeq platform. A color version of the heatmap included in Figure 12A, for example, is available via USPTO PAIR (Accession 63 / 033,176, Additional Content tab).
[0076]
[0109] FIG. 13 is a table of compensatory mutation pairs identified by AlphaSeq relative intensity data based on yeast mating efficiency measured in a per-library screen of pembro scFv (antibody) and PD-1 (antigen). Column 1300 lists the PD-1 mutant protein binding partner, column 1302 lists the paired pembro scFv mutant protein binding partner, and column 1304 lists the minimum distance in angstroms between the paired mutated residues in columns 1300 and 1302. Residue-related compensatory mutation pairs are spatially close within the antibody-antigen interface. The mutant protein binding partner pairs for which relative intensities are plotted in FIGS. 16-17, described below, are highlighted with gray dots.
[0077]
[0110] FIG. 14 is an illustration of the same data set shown in the heat map of FIG. 7, and is a graphical representation of the crystal structure of the antibody-antigen interface. These data show that spatial epitope mapping alone can produce false positive signals. For example, the lysine residue at PD-1 position 107 is spatially close to the antibody as revealed by the distance heat map, and makes a well-defined set of interactions with antibody residues, such as E194 in the VL domain. However, either residue can be ineffectively mutated, and thus the interaction revealed by spatial epitope mapping can be considered a false positive because it is not functionally important, as demonstrated by AlphaSeq mutational analysis. A color version of the heat map, for example included in FIG. 14, is available via USPTO PAIR (accession 63 / 033,176, additional content tab).
[0078]
[0111] FIG. 15 is a diagram illustrating the potential of the AlphaSeq platform to detect compensatory mutations by measuring the relative AlphaSeq signal in a library-by-library screen of pembro scFv (antibody) and PD-1 (antigen). Library-by-library analysis can identify a relatively rare subset of interactions that exhibit the plotted AlphaSeq signal signature. Compensatory mutations that exhibit this signature allow confirmation of wild-type-like mating frequencies for mutant pairs in which at least one, or both, of the mutants have weakened interactions with the cognate wild type. By examining the X-ray structure of the wild-type complex, it has been found that the residues harboring these compensatory mutations are spatially close.
[0079]
[0112] FIG. 16 shows three pairwise plots of mutant protein binding partners showing the signature of compensatory mutations, along with a graphical representation of the crystal structure of the antibody-antigen interface, with relevant residues highlighted.
[0080]
[0113] FIG. 17 shows two pairwise plots of mutant protein binding partners showing the signature of compensatory mutations, along with a graphical representation of the crystal structure of the antibody-antigen interface, with relevant residues highlighted.
[0081]
[0114] Figure 18 is a graphical representation highlighting pairs of compensatory mutations detected by measuring relative AlphaSeq signal in a per-library screen between libraries of pembro scFv (antibody) and PD-1 (antigen) mutants. A total of 10 unique residue-residue interactions are shown, including 7 antibody and 6 antigen residues, with all compensatory pairs spatially close at the antigen / antibody interface. These compensatory mutations provide unique information about the protein-protein interface that is not available using methods based on one-sided binding.
[0082]
[0115] FIG. 19 shows a previous method of epitope mapping by targeted mutagenesis. In the conventional method of epitope mapping known from the past, surface residues on a target protein are mutated one by one on an individual basis to alanine (alanine scanning) or another amino acid, and binding of the target by an antibody is evaluated. In certain embodiments with respect to binding, mutations that disrupt binding of the antibody to the target are predicted to be important and contain the epitope. This approach is slow and expensive because each antibody-target mutant interaction is evaluated separately, or target mutants are batched and epitope mapped one antibody at a time. For example, a target protein 1900 can be subjected to alanine scanning mutagenesis and an epitope mapped to an antibody 1904. Mutation target 1902 contains mutation 1906 of an amino acid residue at position 17 of the protein. Mutation 1906 disrupts binding of mutation target 1902 to antibody 1904, meaning that the epitope of the target is near the amino acid residue at position 17.
[0083]
[0116] FIG. 20 shows a library-by-library screen for epitope mapping using the methods disclosed herein. In some implementations, the target protein 2000 can be subjected to alanine scanning mutagenesis across all amino acid positions of the protein. In other implementations, the target protein 2000 can be subjected to full site saturation mutagenesis, where each amino acid position of the protein is mutated to all available amino acid variants to create a library of mutant target proteins. A library of antibodies, e.g., antibody 2002, is provided and screened against the target protein mutagenesis library according to the methods disclosed herein, and all binding interactions between the target protein library and the antibody library are evaluated. For each antibody in the antibody library, the binding interactions are evaluated and the target protein epitope can be inferred from the position of the mutation that disrupts binding relative to wild-type binding.
[0084]
[0117] FIG. 21 is a graphical representation of the data generated by the per-library screen shown in FIG. 20. The data is shown as a heat map representing the results of a screen of PD-1 variants (antigen variants) against a library of antibodies (one or more). All PD-1 surface residues were selected for mutagenesis and the resulting SSM library was subjected to the AlphaSeq workflow. The AlphaSeq signal is displayed in a heat map format, with various patterns of dark shaded boxes indicating high and low mating frequencies. Ten antibodies were screened against a site-saturation mutagenesis library of PD-1 surface positions. The PD-1 surface positions are shown from left to right along axis 2100, and the test antibodies and controls are shown along axis 2102. A color version of the heat map included in FIG. 21, for example, is available via USPTO PAIR (Accession 63 / 033,176, Additional Content tab).
[0085]
[0118] FIG. 22 is an additional view of two of the antibodies shown in FIG. 21. The data for antibodies 9 and 10 from FIG. 21 have been reconstructed in FIG. 22 as enrichment / depletion heatmaps to show the results of the screens per library. Heatmap 2204 represents data from a screen of pembrolizumab (the antibody) against a library of PD-1 surface residue variants (antigen variants), and heatmap 2206 represents data from a screen of nivolumab (the antibody) against a library of PD-1 surface residue variants (antigen variants). The AlphaSeq signals are displayed in heatmap format, with various patterns of dark shaded boxes indicating high and low conjugation frequencies. The PD-1 surface positions are shown from left to right along axis 2200, and for each PD-1 surface position, the individual amino acid variants are shown along axis 2202. Color versions of the heatmaps included in FIG. 22, for example, are available via USPTO PAIR (Accession 63 / 033,176, Additional Content tab).
[0086]
[0119] FIG. 23 shows a possible output of the method disclosed herein, where compensatory mutations are identified between a first protein binding partner and a second protein binding partner. Wild type protein 2300 and wild type protein 2302 interact as first and second protein binding partners. A full site saturation mutagenesis screen for each library according to the method disclosed herein identified mutations in the first and second protein binding partners at positions 2308, 2310, and 2312 for each mutation such that the mutant protein binding partner interacts in a similar manner to the wild type interaction, but the mutant protein binding partner does not interact with the wild type protein binding partner. For example, mutant protein binding partner 2304 interacts strongly with mutant protein binding partner 2306, but does not interact with wild type protein binding partner 2302 due to mutations at positions 2308, 2310, and 2312. The identification and design of such orthogonal protein-binding interactions may be useful for applications such as engineering antibodies, engineering synthetic receptor / ligand pairs, or for synthetic biology tools such as engineered enzyme scaffolds.
[0087]
[0120] Figure 24A shows a library-by-library screen for epitope mapping using the methods disclosed herein. A library-by-library screen was performed between a site-saturation mutagenesis library of PD-1 (antigen) surface residue variants and a site-saturation mutagenesis library of pembrolizumab (antibody) variants. 19 amino acid positions of PD-1 were selected for mutagenesis, and 33 amino acid positions of pembrolizumab were selected for mutagenesis. Amino acid residues of the antigen and antibody were selected near the protein-protein binding interface. The AlphaSeq platform was used to measure all pairwise interactions between the site-saturation mutagenesis library of antigen and antibody protein binding partners, including over 220,000 pairwise interactions. Heatmap 2400 illustrates pairwise library-by-library intensity measurements by AlphaSeq of the interaction of a library of PD-1 mutants with a library of pembrolizumab mutants relative to the binding affinity of wild-type PD-1 and wild-type pembrolizumab, with lighter shading representing wild-type binding affinity and darker shading representing reduced binding affinity.
[0088]
[0121] Figure 24B shows a more detailed view of a subset of the data shown in heatmap 2400 of Figure 24A. Heatmap inset 2402 represents pairwise interaction data between 20 PD-1 variants with mutations at position 54 and 20 pembrolizumab variants with mutations at position 54, or 400 total protein-protein interaction data measured by the methods disclosed herein. In a single AlphaSeq assay, binding affinity data was measured for all pairwise combinations of 33 select pembrolizumab positions with 19 select PD-1 positions and assessed against wild-type binding affinity between the two protein binding partners.
[0089]
[0122] Figure 24C highlights specific pairwise interactions between single PD-1 mutations and single pembrolizumab variants. Heatmap 2412 illustrates AlphaSeq pairwise library-by-library intensity measurements of interactions between libraries of PD-1 and pembrolizumab variants versus binding affinity of wild-type PD-1 and wild-type pembrolizumab, with light shading representing wild-type binding affinity and dark shading representing reduced binding affinity, highlighting two specific mutations: PD-1K54F and pembrolizumab Y33P. Box 2404 represents binding affinity for wild-type PD-1 and wild-type pembrolizumab, and is white according to the heatmap shading. Box 2406 represents binding affinity of wild-type PD-1 with pembrolizumab Y33P, and is darkly shaded, indicating significantly reduced binding affinity relative to wild-type. Box 2408 represents the binding affinity of PD-1K54F with wild type pembrolizumab and is shaded dark, indicating a significantly reduced binding affinity relative to wild type. Box 2410 represents the binding affinity of PD-1K54F with pembrolizumab Y33P and is shaded white, indicating a binding affinity similar to wild type. That is, with respect to these individual mutations, each mutation by itself significantly reduces the binding affinity between the first and second protein binding partners, but when the mutations are present simultaneously, the binding affinity is restored to a level similar to the wild type binding affinity. The mutation in the first protein binding partner compensates for the binding defect caused by the mutation in the second protein binding partner, promoting a binding affinity similar to the wild type binding affinity. Color versions of the heat maps included in, for example, Figures 24A-24C are available via USPTO PAIR (Accession 63 / 033,176, Additional Content tab).
[0090]
[0123] Figure 24D is a graphical representation of the data shown in heatmap 2412 of Figure 24C. Four pairwise interactions of wild-type and mutant PD-1 in combination with pembrolizumab are shown. The graph is normalized such that the binding affinity of wild-type PD-1 with wild-type pembrolizumab is set to 1.0 and pairwise interactions of mutant protein-binding partners are quantified relative to 1.0. As described with respect to Figure 24C, these mutations exhibit the unique and unexpected property of compensating for the deleterious effect on binding affinity that each mutation exerts by itself, such that PD-1K54F and pembrolizumab Y33P exhibit binding similar to the binding affinity between the wild-type protein-binding partners.
[0091]
[0124] FIG. 25 shows a first plot of expected and measured interaction strengths between two protein binding partners, and a second plot of the measured interaction strengths versus expected interaction strengths between antibody-antigen protein binding partners assessed using the methods disclosed herein. The expected interaction strength may be defined by multiplying the relative binding affinity of the mutant first protein binding partner to the wild type by the relative binding affinity of the mutant second protein binding partner to the wild type. That is, the expected interaction strength of the interaction between PD-1K54F and pembrolizumab Y33P may be defined by multiplying the relative affinity of PD-1K54F to wild type pembrolizumab by the relative affinity of pembrolizumab Y33P to wild type PD-1. As shown in plot 2500, the expected interaction strength of PD-1K54F and pembrolizumab Y33P is near zero because the binding affinity of each individual mutation with its corresponding wild type protein binding partner is substantially reduced. However, the measured interaction strength between PD-1K54F and pembrolizumabY33P is almost the same as that between wild-type PD-1 and wild-type pembrolizumab due to the unexpected compensation effect of these mutations. Plot 2502 shows the expected interaction strength plotted against the measured interaction strength, and indicates and highlights the compensatory mutations in light gray. These compensatory mutations are pairs of mutant protein binding partners whose measured interaction strength significantly exceeds the expected interaction strength due to the unexpected compensation effect of the mutation.
[0092]
[0125] Figure 26 is a plot of the ratio of measured interaction strength to the distance between amino acid residues between protein binding partners and expected interaction strength to the distance between amino acid residues between protein binding partners. Compensatory mutations whose measured interaction strength is significantly higher than expected interaction strength are highlighted in light gray. This plot demonstrates that pairs of amino acid residues identified as compensatory mutations are physically close to each other at the protein-protein interface. All compensatory mutations identified by the method disclosed herein are less than 7 angstroms apart at the protein-protein interface according to known X-ray crystal structures.
[0093]
[0126] Figure 27 is a 3D model based on the X-ray crystal structure of the interface between PD-1 (antigen) and pembrolizumab (antibody). The amino acid residues with compensatory mutations identified are highlighted in light grey. This model demonstrates that all of the amino acid residues identified as compensatory mutations are physically close to each other at the protein-protein interface.
[0094]
[0127] In some implementations, the invention provides a novel method for identifying compensatory mutations between two protein binding partners, comprising: providing a first library of protein binding partners comprising a first wild-type polypeptide and a first plurality of mutant polypeptides; providing a library of second protein binding partners comprising a second wild-type polypeptide and a second plurality of mutant polypeptides; determining affinity measurements between each protein binding partner of the library of first protein binding partners and each protein binding partner of the library of second protein binding partners; based on the affinity measurements of each protein binding partner of the first library of protein binding partners and each protein binding partner of the second library of protein binding partners, (i) a polypeptide of a first plurality of mutant polypeptides, and (ii) one polypeptide of the second plurality of mutant polypeptides; identifying one or more pairs of protein binding partners comprising: the measured affinity of each of the one or more pairs of protein binding partners is substantially different from the respective expected affinity values between each respective pair of protein binding partners; The affinity expectation for a given pair of protein binding partners is a) an affinity measurement between a first wild-type polypeptide of a given pair and one polypeptide of a second plurality of mutant polypeptides of a given pair; and b) an affinity measurement between one polypeptide of a first plurality of mutant polypeptides of a given pair and a second wild-type polypeptide of a given pair; The present invention provides a method for calculating the above-mentioned
[0095]
[0128] In some implementations, each protein binding partner of the first library of protein binding partners is expressed on the surface of one of the first plurality of yeast cells, and each protein binding partner of the second library of protein binding partners is expressed on the surface of one of the second plurality of yeast cells.
[0096]
[0129] In some implementations, the affinity measurements for each protein binding partner of the first library of protein binding partners and each protein binding partner of the second library of protein binding partners are measured by synthetic agglutination of the first plurality of yeast cells with the second plurality of yeast cells.
[0097]
[0130] In some implementations, each protein binding partner of the first library of protein binding partners is an antibody, scFv, Fab, or VHH species.
[0098]
[0131] In some implementations, each protein binding partner of the second library of protein binding partners is an antigen species.
[0099]
[0132] In some implementations, each protein binding partner of the first library of protein binding partners is a receptor species.
[0100]
[0133] In some implementations, each protein binding partner of the second library of protein binding partners is a ligand species.
[0101]
[0134] In some implementations, each of the first plurality of mutant polypeptides and each of the second plurality of mutant polypeptides are produced by user-directed mutagenesis.
[0102]
[0135] In some implementations, the measured affinity of each of one or more pairs of protein-binding partners is substantially higher than the expected affinity of the protein-binding partner pair.
[0103]
[0136] In some implementations, the measured affinity of each of one or more pairs of protein binding partners is more than twice as high as the expected affinity of the protein binding partner pair.
[0104]
[0137] In some implementations, the measured affinity of each pair of one or more pairs of protein binding partners is substantially lower than the expected affinity of the pair of protein binding partners.
[0105]
[0138] In some implementations, the measured affinity of each of one or more pairs of protein-binding partners is more than two-fold lower than the expected affinity of the protein-binding partner pair.
[0106]
[0139] In some implementations, the invention provides a novel method for identifying compensatory mutations between two protein binding partners, comprising: providing a first library of protein binding partners comprising a first wild-type polypeptide and a first plurality of mutant polypeptides; providing a second library of protein binding partners comprising a second wild-type polypeptide and a second plurality of mutant polypeptides; determining affinity measurements between each protein binding partner of the first library of protein binding partners and each protein binding partner of the second library of protein binding partners; identifying one or more pairs of protein binding partners having respective affinity measurements that are substantially different from the affinity measurement between the first wild-type polypeptide and the second wild-type polypeptide based on the affinity measurements between each protein binding partner in the first library of protein binding partners and each protein binding partner in the second library of protein binding partners; The present invention provides a method comprising:
[0107]
[0140] In some implementations, one or more pairs of protein binding partners meet the following conditions: a. the measured affinity between one polypeptide of the first plurality of mutant polypeptides and the second wild-type polypeptide is substantially less than the measured affinity between the first wild-type polypeptide and the second wild-type polypeptide; b. the measured affinity between the first wild-type polypeptide and one of the second plurality of mutant polypeptides is substantially less than the measured affinity between the first wild-type polypeptide and the second wild-type polypeptide; and c. the affinity measurement between one polypeptide of the first plurality of mutant polypeptides and one polypeptide of the second plurality of mutant polypeptides is substantially the same as or substantially higher than the affinity measurement between the first wild-type polypeptide and the second wild-type polypeptide; Meet the following.
[0108]
[0141] In some implementations, one or more pairs of protein binding partners meet the following conditions: a. the measured affinity between one polypeptide of the first plurality of mutant polypeptides and the second wild-type polypeptide is substantially the same as or substantially higher than the measured affinity between the first wild-type polypeptide and the second wild-type polypeptide; b. the measured affinity between the first wild-type polypeptide and one of the second plurality of mutant polypeptides is substantially less than the measured affinity between the first wild-type polypeptide and the second wild-type polypeptide; and c. the affinity measurement between a polypeptide of the first plurality of mutant polypeptides and a polypeptide of the second plurality of mutant polypeptides is substantially the same as the affinity measurement between the first wild-type polypeptide and the second wild-type polypeptide; Meet the following.
[0109]
[0142] In some implementations, one or more pairs of protein binding partners meet the following conditions: a. the measured affinity between one polypeptide of the first plurality of mutant polypeptides and the second wild-type polypeptide is substantially less than the measured affinity between the first wild-type polypeptide and the second wild-type polypeptide; b. the measured affinity between the first wild-type polypeptide and one of the second plurality of mutant polypeptides is substantially the same as or substantially higher than the measured affinity between the first wild-type polypeptide and the second wild-type polypeptide; and c. the measured affinity between a polypeptide of the first plurality of mutant polypeptides and a polypeptide of the second plurality of mutant polypeptides is substantially the same as or substantially greater than the measured affinity between the first wild-type polypeptide and the second wild-type polypeptide; Meet the following.
[0110]
[0143] In some implementations, one or more pairs of protein binding partners meet the following conditions: a. the measured affinity between one polypeptide of the first plurality of mutant polypeptides and the second wild-type polypeptide is substantially the same as or substantially higher than the measured affinity between the first wild-type polypeptide and the second wild-type polypeptide; b. the measured affinity between the first wild-type polypeptide and one of the second plurality of mutant polypeptides is substantially the same as or substantially higher than the measured affinity between the first wild-type polypeptide and the second wild-type polypeptide; and c. the measured affinity between a polypeptide of the first plurality of mutant polypeptides and a polypeptide of the second plurality of mutant polypeptides is substantially less than the measured affinity between the first wild-type polypeptide and the second wild-type polypeptide; Meet the following.
[0111]
[0144] In some implementations, one or more pairs of protein binding partners meet the following conditions: a. the measured affinity between a polypeptide of the first plurality of mutant polypeptides and a second wild-type polypeptide, or between the first wild-type polypeptide and a polypeptide of the second plurality of mutant polypeptides, is substantially lower than the measured affinity between the first wild-type polypeptide and the second wild-type polypeptide; b. the measured affinity between a polypeptide of the first plurality of mutant polypeptides and a polypeptide of the second plurality of mutant polypeptides is substantially greater than the measured affinity between a polypeptide of the first plurality of mutant polypeptides and the second wild-type polypeptide or the measured affinity between the first wild-type polypeptide and a polypeptide of the second plurality of mutant polypeptides; Meet the following.
[0112]
[0145] In some implementations, mutations in one polypeptide of the first plurality of mutant polypeptides define a paratope of an antibody, scFv, Fab, or VHH species, and / or mutations in one polypeptide of the second plurality of mutant polypeptides define an epitope of an antigen species.
[0113]
[0146] In some implementations, mutation of one polypeptide of the first plurality of mutant polypeptides and mutation of one polypeptide of the second plurality of mutant polypeptides. a. one polypeptide of a first plurality of mutant polypeptides binds to a second wild-type polypeptide and to one polypeptide of a second plurality of mutant polypeptides; b. An orthogonal binding relationship is created between a polypeptide of the first plurality of mutant polypeptides and a polypeptide of the second plurality of mutant polypeptides such that a polypeptide of the second plurality of mutant polypeptides binds to a polypeptide of the first plurality of mutant polypeptides and does not bind to a second wild-type polypeptide.
[0114]
[0147] In some implementations, mutation of one polypeptide of the first plurality of mutant polypeptides and mutation of one polypeptide of the second plurality of mutant polypeptides. a. one polypeptide of the first plurality of mutant polypeptides binds to one polypeptide of the second plurality of mutant polypeptides and does not bind to the second wild-type polypeptide; b. An orthogonal binding relationship is created between a polypeptide of the first plurality of mutant polypeptides and a polypeptide of the second plurality of mutant polypeptides such that a polypeptide of the second plurality of mutant polypeptides binds to a polypeptide of the first plurality of mutant polypeptides and does not bind to the first wild-type polypeptide.
[0115]
[0148] In some implementations, affinity binding measured by the methods disclosed herein may be output to a digital display device. In another implementation, the numerical and graphical representations of affinity binding data for wild-type and mutant protein-binding partners measured by the methods disclosed herein may be presented on a display device with notation indicating pairs of mutant protein-binding partners with mutations that have been identified as compensatory mutations.
[0116]
[0149] In some implementations, for mutant protein binding partners with one or more mutations identified as compensatory mutations by the methods disclosed herein, mutations can be used to engineer protein interactions with orthogonal binding affinities as detailed above. For example, in some implementations, compensatory mutations identified by the methods disclosed herein can be used to construct engineered metabolic pathways that include enzymes heterologous to the production host organism, for example, to produce useful secondary metabolites, and the interactions and titers of pathway component enzymes can be fine-tuned by the use of compensatory mutations. Furthermore, heterologous metabolic pathway components can be engineered using compensatory mutations identified by the methods disclosed herein to not interact with proteins and enzymes in the host organism that would otherwise impair or reduce the activity of the heterologous metabolic pathway.
[0117]
[0150] In another implementation, compensatory mutations for a receptor can be identified by the methods disclosed herein, allowing engineering of CAR-T cells to express customized cell surface receptors with compensatory mutations. Similarly, a soluble growth factor or cytokine can be engineered to express compensatory mutations identified by the methods disclosed herein, such that the CAR-T cell surface receptor and the soluble growth factor or cytokine exhibit a one-sided orthogonal affinity relationship. By introducing a possibly very small number of highly impactful compensatory mutations into each of the cell surface receptor and the soluble growth factor or cytokine, the CAR-T cell surface receptor can bind and be activated by both the engineered growth factor or cytokine and the native wild-type growth factor or cytokine to the patient's physiological environment. Conversely, the engineered soluble growth factor or cytokine with compensatory mutations identified by the methods disclosed herein binds and activates only the engineered CAR-T cell surface receptor, and does not affect the native multiple or wild-type cell surface receptors to the patient's physiology. This pattern of customized orthogonal protein-protein interactions using the methods disclosed herein will be useful in engineering cell therapies, immunotherapies, and biologics to treat a number of diseases and disorders.
[0118]
[0151] In another implementation, the compensatory mutations identified by the methods disclosed herein can be useful for the rational design of antibody-based immunotherapies. In one implementation, an antibody, scFv, Fab, or VHH species can be engineered to carry the compensatory mutations identified by the methods disclosed herein such that its affinity and properties for its antigen are tunable and customizable. In another implementation, an antibody, scFv, Fab, or VHH species can be engineered to carry the compensatory mutations identified by the methods disclosed herein such that the antibody, scFv, Fab, or VHH species specifically binds to a novel epitope that is different from the epitope of the wild-type antibody, scFv, Fab, or VHH species. EXAMPLES
[0119] Example 1
[0152] Using the AlphaSeq platform (see, e.g., Example 2, and also U.S. Patent Publication No. 2017-0205421A1) and the methods disclosed herein, a library of mutants of programmed cell death protein 1 (PD-1), a cell surface receptor expressed on T cells and pro-B cells, was screened against a library of mutants of the short chain variable fragment (scFv) of the monoclonal antibody pembrolizumab (Pembro), a humanized antibody used in cancer immunotherapy for the treatment of melanoma, lung cancer, and Hodgkin's lymphoma, among other cancers. The library of Pembro scFv mutants included a comprehensive site-saturation mutagenesis library of 33 amino acid residues spanning several domains from positions 30 to 235 of the polypeptide. The library of PD-1 mutants included a comprehensive site-saturation mutagenesis library of 60 amino acid residues spanning several domains from positions 5 to 115 of the polypeptide. The AlphaSeq screen for each library allowed for the interrogation of the affinity of each PD-1 mutant with each Pembro scFv mutant in a pairwise manner.
[0120]
[0153] The above experiment of screening the PD-1 mutant library against wild-type pembro scFv, the results of which are shown in FIG. 7, identified PD-1 residue K54 as particularly intolerant of amino acid substitutions, as reflected by the widespread low conjugation frequency of amino acid substitutions at that position. The library-by-library screen of PD-1 and pembro scFv mutants identified a subset of pairs of compensatory mutations of the two protein binding partners. The results of the subset of compensatory mutations are plotted in FIG. 16. For example, the affinity of the PD-1 mutant K54F with wild-type pembro scFv was 0.05 relative to the wild type due to the wild-type interaction of the two protein binding partners (n=3; standard deviation=0.03). The affinity of the pembro scFv mutant Y33P with wild-type PD-1 was 0.33 relative to the wild type due to the wild-type interaction of the two protein binding partners (n=3; standard deviation=0.08). However, the affinity of PD-1 mutant K54F with pembro scFv mutant Y33P was 1.19 relative to wild type with wild type interaction of the two protein binding partners (n=3; standard deviation=0.38), i.e., the interaction of these two mutants of protein binding partners was nearly equivalent to wild type with wild type interaction, indicating a pair of compensatory mutations of the two protein binding partners. Similar signatures of affinity between the two binding partners were identified for PD-1 mutant K54M and pembro scFv mutant Y33P. For the third pair of mutant protein-binding partners, the affinity of the PD-1 mutation K54I with wild-type pembro scFv was 0.40 relative to the wild-type with the wild-type interaction of the two protein-binding partners (n=3; standard deviation=0.22), and the affinity of the pembro scFv mutation Y101K with wild-type PD-1 was 1.06 relative to the wild-type with the wild-type interaction of the two protein-binding partners (n=3; standard deviation=0.18), meaning that the pembro scFv mutation Y101K did not affect affinity.However, the affinity of the PD-1 mutant K54I with the pembro scFv mutant Y101K was 1.87 (n=3; standard deviation=0.35) relative to the wild type for the wild type interaction of the two protein binding partners, i.e., the interaction of these two mutants of the protein binding partners had a significantly higher affinity than the wild type for the wild type interaction of the two protein binding partners, indicating a pair of compensatory mutations of the two protein binding partners. The compensatory mutations identified by these experiments correspond to amino acid residues of the antigen and antibody that are spatially close to the antigen / antibody interface.
[0121] Example 2
[0154] Development of the (AlphaSeq) Yeast-Mating Assay for Screening and / or Determining Protein-Protein Interactions and Protein Interaction Networks
[0155] Flow cytometric assays can be used to distinguish between MATa, MATalpha and diploid cells. The native yeast sexual agglutinin has been replaced with a surface-displayed binder (SAP) and mating efficiency was measured using flow cytometry. A diploid chromosome translocation system was developed to combine genes for both binders on a single chromosome so that next generation sequencing could be used to assess the mating frequency of a particular pair of binders in large libraries.
[0122]
[0156] Although there are many cell-based assays to analyze extracellular binding between libraries of proteins and single targets, only cell-free approaches have been developed to characterize the entire protein interaction network in a single assay. This implies a time-consuming and expensive library generation process that requires purification and labeling of each protein component in the network. This example demonstrates a paired yeast surface display (PYSD) assay for library-on-library characterization of protein interactions that combines yeast surface display and sexual agglutinins to couple protein binding to mating in S. cerevisiae. Notably, this example demonstrates that sexual agglutinins are highly engineerable by knocking out the native agglutinin protein and instead displaying a complementary binding protein (synthetic agglutinin, SAP) on the surface of MATa and MATalpha yeast cells. This example shows that mating efficiency is highly dependent on the binding affinity and expression level of the surface-expressed protein. The chromosomal translocation scheme allows the protein-protein interaction network to be analyzed by next generation sequencing and applied to the analysis of two engineered protein interaction networks.
[0123]
[0157] Characterization of protein interaction networks, both binding affinity and specificity, is important for understanding cellular functions, screening therapeutic candidates, and evaluating engineered protein networks. For example, protein "interactome" mapping has expanded our understanding of biological systems and disease states, and can be used to evaluate therapeutic candidates for appropriate mediation or disruption of specific protein interactions. Furthermore, the construction of synthetic systems often requires highly specific and orthogonal protein interactions to adequately control cellular behavior. Engineered protein binding domains that enable the construction of arbitrary protein interaction networks need to be carefully characterized in the context of highly complex biological systems.
[0124]
[0158] There are many approaches to analyze the binding of libraries of proteins to single protein targets. Yeast surface display (YSD) is widely used, in part because of the ease of library construction. However, to analyze protein networks, it is necessary to screen the binding between all possible pairs of proteins. Because YSD measures binding to cellular fluorescence after incubation with a soluble fluorescently labeled target, this approach does not allow screening against libraries of target proteins. Recently developed approaches use DNA barcoded proteins for one-pot library-on-library characterization, but require purification of each constituent protein in the network, making analysis of large networks time-consuming and expensive. This disclosure presents a novel method that combines the ease of YSD library creation with a high-throughput assay that can characterize entire protein interaction networks in a single pot.
[0125]
[0159] Paired yeast surface display (PYSD) platforms are used for one-to-one, many-to-one, or many-to-many protein interaction characterization. For the one-to-one screen, two isogenic display strains, one MATa constitutively expressing a fluorescent marker (e.g., mCherry) and one MATalpha constitutively expressing a second fluorescent marker (e.g., mTurquoise), each expressing a synthetic adhesion protein (SAP) on its surface as a fusion to Aga2 (Aga2-myc). A mating assay is then used to determine the effect of displaying those particular SAPs on mating efficiency, reported as the percentage of diploid cells after 17 hours. Haploids and diploids are differentiated based on their expression of mCherry and mTurquoise in a flow cytometry assay. The surface expression intensity of each SAP is determined by incubating the mixed culture with a FITC-conjugated anti-myc antibody prior to flow cytometry.
[0126]
[0160] For many-to-one or many-to-many screens, one or both of the isogenic displayer strains are replaced with a display library or a library of displayer cells, each expressing a unique SAP. After a short mating period, the cells are transferred to a medium lacking lysine and leucine, which is used to select for diploid cells only. For many-to-many screens, β-estradiol (βE) is also added to induce CRE recombinase expression in mating diploids. Recombinase expression results in translocation at the lox66 / lox71 sites, which flanks the SAP integration and brings the SAP gene proximal to one copy of chromosome III. Due to the biased nature of the lox66 / 71 recombinase site pairs, the majority of the population consists of translocation diploids. After translocation, the SAP pairs are no longer uncoupled from certain diploid cells by cell lysis. For both many-to-one and many-to-many screens, colony PCR of the diploid populations is analyzed by next generation sequencing to determine the mating frequency of each SAP compared to all other potential SAPs included in the assay.
[0127] Materials and Methods:
[0161] Plasmid construction: The plasmids used in the first example (Example A) are shown in Table 1. For each construct, the backbone and insert were amplified by PCR, gel extracted, and assembled into a plasmid using a Gibson reaction. Standard linkers between all parts increased the efficiency and consistency of cloning. All backbones, consisting of a high copy origin of replication and ampicillin resistance, were flanked with Pme1 restriction sites for easy linearization and integration into the yeast chromosome. All plasmids contain approximately 500 bases of chromosomal homology upstream and downstream of the target locus. Knockout (KO) plasmids contain upstream and downstream chromosomal homology but no gene cassette. The sequences of each promoter, open reading frame, terminator, and chromosomal homology were verified by Sanger sequencing.
[0128] [Table 1]
[0129]
[0162] Yeast strain generation and growth conditions: The S. cerevisiae strains used in the second example (Example B) are shown in Table 2. EBY100a and W303αMOD were used as the initial parent strains. EBY100α was produced by mating these two parent strains, followed by sporulation and screening in tetrads for the appropriate selection marker. All other strains were generated with chromosomal integration by linearizing the given plasmid with Pme1 restriction digestion and performing a standard LiAc transformation procedure. Selection of transformants was achieved using media lacking the given auxotrophic marker or using media supplemented with eukaryotic antibiotics. Following each transformation, diagnostic colony PCR was performed to verify integration at the appropriate locus. All yeast assays used standard yeast culture media and were grown at 30°C. All liquid culture growth was performed in 3 mL of YPD liquid medium with shaking at 275 RPM.
[0130] [Table 2]
[0131] [Table 3]
[0132]
[0163] Mating assay: To evaluate the mating efficiency between any two yeast strains in liquid culture, haploid strains were first grown to saturation or approximately 18 hours from isogenic colonies on fresh YPD plates. Each haploid was then combined in a fresh 3mL YPD liquid culture such that the MATa strain was at a density of 100 cells / μL and the MATα strain was at a density of 600 cells / μL. This difference in starting concentrations was an adjustment for the confirmed unequal growth response to mating factors. Cells were also grown separately in fresh YPD to individually evaluate their surface expression intensity. After 17 hours of growth, 2.5μL of mating culture was added to 1mL of molecular grade water and read on a flow cytometer. MATa, MATalpha, and diploid cells were differentiated based on the fluorescence intensity of mCherry and mTurquoise. Miltenyi MACSQUANT® VYB was used in the above experiments. The Y2 channel (561 nm excitation laser and 615 nm emission filter) was used to measure mCherry expression, and the V1 channel (405 nm excitation laser and 450 nm emission filter) was used to measure mTurquoise expression. The diploid cell population as a percentage of the total cell population after 17 hours was used as a measure of mating efficiency. Surface expression intensity was measured by incubating 10 μL of each individually grown cell line with FITC-conjugated anti-myc antibody in PBSF for 15 minutes after washing in 1 mL water. The cells were then washed again and resuspended in 1 mL water. Flow cytometry was then performed. For the determination of surface expression intensity, an ACCURI™ C6 cytometer was used. FITC binding to the cells was measured using the FL1.A channel (488 nm excitation laser and 533 nm emission filter). FLOWJO™ was used for all cytometric analyses.
[0133]
[0164] For one-to-many batch mating assays, recombinant MATa yeast strains expressing a single SAP fused to Aga2 are combined with multiple recombinant MATalpha yeast strains expressing different SAPs fused to Aga2 in a fresh 3 mL YPD culture. MATa strains are added at a density of 100 cells / μL, and MATalpha strains are added at equal concentrations for a total concentration of 600 cells / μL. After 6 hours of growth, hygromycin is added at 100 ng / μL. 20 hours after inoculation of the initial culture, 1 mL of cells are pelleted. 2 μL of cells are removed from the pellet, lysed in 0.2% SDS, spun down to remove all cellular debris, and diluted in water. The lysate is then used as template for PCR with standard primers containing overhangs for next-generation sequencing, and the PCR product, expected to be approximately 350 bases, is purified from the gel slice. Single-read next-generation sequencing was then performed. The frequency with which a particular barcode is identified relative to the total number of reads provides a relative measure of the number of junctions generated by the SAP associated with that particular barcode.
[0134]
[0165] For many-to-many (see also Example 6) batch mating assays, multiple haploid yeast strains of each mating type are combined in a fresh 3 mL YPD culture. Recombinant MATa yeast strains are added at equal concentrations for a total density of 100 cells / μL, and recombinant MATalpha yeast strains are added at equal concentrations for a total density of 600 cells / μL. After 6 hours of growth, hygromycin is added at 100 ng / μL, and β-estradiol (βE) is added at 200 ng / μL. 20 hours after inoculation of the initial culture, 1 mL of cells are pelleted. 2 μL of cells are removed from the pellet, lysed in 0.2% SDS, spun down to remove all cellular debris, and diluted in water. The lysate is then used as template for PCR with standard primers containing overhangs for next-generation sequencing, and the PCR product, expected to be 650 bases, is purified from the gel slice. Paired-end next-generation sequencing was then performed. The frequency with which a particular pair of barcodes is identified relative to the total number of reads provides a relative measure of the number of junctions made by the SAP pair associated with two particular barcodes.
[0135] result
[0166] For S. cerevisiae haploid cells lacking essential sexual agglutinin proteins, conjugation is sufficient to restore aggregation and mating in liquid culture. Sag1, the primary MATalpha sexual agglutinin protein, is essential for aggregation. When Sag1 is knocked out, MATalpha cells are unable to mate with wild-type MATa cells in turbulent liquid culture. However, complementary SAPs expressed on the display pair are able to restore mating. Non-complementary SAPs are unable to restore mating.
[0136]
[0167] The frequency of mating events between any two display cells depends on the binding affinity between the SAP pair and the surface expression intensity of each SAP. The results demonstrate that the binding affinity and the measured mating efficiency are positively correlated. However, it is possible to improve the correlation by adjusting the mating efficiency for the expression level of each SAP.
[0137]
[0168] Seven SAP pairs with known affinity were evaluated for mating efficiency (see Table 3). The mating efficiency of each pair was tested four times, and the average and standard deviation were calculated. The surface expression intensity (SES) of each haploid display strain was also measured, as described in Materials and Methods. A "mating score," which adjusts mating efficiency for differences in surface expression intensity, was calculated by dividing the average mating efficiency by the product of the surface expression intensities of both haploid display strains.
[0138] [Table 4]
[0139]
[0169] Batch mating allows the determination of relative interactions between many proteins in a single assay. By barcoding each SAP, a many-to-one screen allows the relative mating frequency of a particular SAP with an SAP library to be assessed using single-read next-generation sequencing. A CRE recombinase-based translocation scheme can be used to align barcodes from each mating type on the same chromosome. Adding this chromosomal translocation step allows the relative mating frequency between two SAP libraries to be assessed using paired-end next-generation sequencing. This approach allows the analysis of any protein interaction topology.
[0140]
[0170] Further mating assays can be performed with more SAP pairs and with all the measured affinities of the SAP pairs to provide the surface expression intensity and mating efficiency of each SAP. Statistical analysis can be performed to determine the relationship between the binding affinity of each SAP, the surface expression intensity, and the mating efficiency. The results can be used to determine the predictive relationship between these four variables, so that the mating efficiency and the surface expression intensity can be used to estimate the binding affinity and to determine the threshold for detectable recovery of mating efficiency.
[0141]
[0171] Example 2 demonstrates a paired yeast surface display assay that allows library-on-library characterization of protein interactions in a single assay. By replacing the native S. cerevisiae agglutinin protein with a synthetic adhesion protein, it is possible to match mating efficiency and protein binding strength. This approach can then be used to assess binding between two specific proteins or to determine the relative interaction strength between libraries of proteins.
[0142]
[0172] Although some embodiments have been described, these embodiments are presented by way of example only and are not intended to limit the scope of the present disclosure. Indeed, the novel methods, apparatus, and systems described herein may be embodied in other various forms. Furthermore, various omissions, substitutions, and changes in the form of the methods, apparatus, and systems described herein may be made without departing from the spirit of the present disclosure. The appended claims and their equivalents are intended to cover any form or modification that falls within the scope and spirit of the present disclosure.
Claims
1. 1. A method for identifying compensatory mutations between two protein binding partners, comprising: providing a first library of first protein binding partners expressed on the surface of a first plurality of haploid yeast cells, the first library of first protein binding partners comprising a first wild-type polypeptide and a first plurality of mutant polypeptides of the first wild-type polypeptide; providing a second library of second protein binding partners expressed on the surface of a second plurality of haploid yeast cells, the second library of second protein binding partners comprising a second wild-type polypeptide and a second plurality of mutant polypeptides of the second wild-type polypeptide; culturing the first and second populations of haploid yeast cells such that if the first and second protein binding partners interact, diploid yeast cells are produced; determining an affinity measurement between each protein binding partner of the first library of first protein binding partners and each protein binding partner of the second library of second protein binding partners; and identifying said compensatory mutations based on affinity measurements between each first protein binding partner and each second protein binding partner, wherein one or more pairs of said first and second protein binding partners have an affinity measurement that is substantially different from an affinity measurement between said first wild-type polypeptide and said second wild-type polypeptide; the measured affinity of each of said one or more pairs of protein-binding partners is substantially different from the respective expected affinity values between said respective pairs of protein-binding partners; The affinity expectation for a given pair of protein binding partners is (1) normalizing the binding affinity for each wild type to 1.0; (2) calculating the relative binding affinity of each of the mutant protein binding species interactions with its wild-type protein binding partner to obtain a first mutant protein binding affinity and a second mutant protein binding affinity; and (3) multiplying the first mutant protein binding affinity and the second mutant protein binding affinity to obtain an expected binding affinity for the interaction of the two protein-binding partners with each other.
2. a. the measured affinity between a polypeptide of the first plurality of mutant polypeptides and the second wild-type polypeptide is substantially lower than the measured affinity between the first wild-type polypeptide and the second wild-type polypeptide; b. the measured affinity between the first wild-type polypeptide and one of the second plurality of mutant polypeptides is substantially the same as or substantially lower than the measured affinity between the first wild-type polypeptide and the second wild-type polypeptide; and c) The method of claim 1, wherein the affinity measurement between a polypeptide of the first plurality of mutant polypeptides and a polypeptide of the second plurality of mutant polypeptides is substantially the same as or substantially higher than the affinity measurement between the first wild-type polypeptide and the second wild-type polypeptide.
3. a. the measured affinity between a polypeptide of the first plurality of mutant polypeptides and the second wild-type polypeptide is substantially greater than the measured affinity between the first wild-type polypeptide and the second wild-type polypeptide; b. the measured affinity between the first wild-type polypeptide and one of the second plurality of mutant polypeptides is substantially the same as or substantially greater than the measured affinity between the first wild-type polypeptide and the second wild-type polypeptide; and c) The method of claim 1, wherein the affinity measurement between a polypeptide of the first plurality of mutant polypeptides and a polypeptide of the second plurality of mutant polypeptides is substantially the same as or substantially lower than the affinity measurement between the first wild-type polypeptide and the second wild-type polypeptide.
4. a. the measured affinity between a polypeptide of the first plurality of mutant polypeptides and the second wild-type polypeptide is substantially the same as the measured affinity between the first wild-type polypeptide and the second wild-type polypeptide; b. the measured affinity between the first wild-type polypeptide and one of the second plurality of mutant polypeptides is substantially the same as the measured affinity between the first wild-type polypeptide and the second wild-type polypeptide; and c) The method of claim 1, wherein the measured affinity between a polypeptide of the first plurality of mutant polypeptides and a polypeptide of the second plurality of mutant polypeptides is substantially higher or lower than the measured affinity between the first wild-type polypeptide and the second wild-type polypeptide.
5. 2. The method of claim 1, wherein the mutations in one polypeptide of the first plurality of mutant polypeptides define a paratope of an antibody, scFv, Fab, or VHH species.
6. The method of claim 1 , wherein each protein binding partner of said second library of second protein binding partners is an antigen species.
7. The method of claim 1 , wherein each protein binding partner of said first library of first protein binding partners is a receptor species.
8. The method of claim 1 , wherein each protein binding partner of said second library of second protein binding partners is a ligand species.
9. The method of claim 1 , wherein each of the first plurality of mutant polypeptides and each of the second plurality of mutant polypeptides are produced by user-directed mutagenesis.
10. The method of claim 1 , wherein three or more compensatory mutations are identified.
Citation Information
Patent Citations
Riboswitches, methods and compositions for using and in conjunction with riboswitches.
JP2010528617A
